There is a sentence I have written more times than any other consequential sentence in this work, and it has never once gotten easier to write. In one variation or another it reads: the evidence does not permit a finding, on a balance of probabilities, that the alleged conduct occurred. On the page it looks like neutrality itself — careful, measured, exactly as unassuming as a finding is supposed to be. Then you deliver it, in two meetings that are sometimes on the same afternoon, and you watch the same sentence land as an injury on both sides of a table. The complainant hears we did not believe you. The respondent hears not guilty — but not innocent either, and understands the question will trail them anyway. Two people, one finding, and both walk out convinced the process failed them. I am writing at the level of pattern here, as I always will in this space — this is not the story of one difficult case. It is what the end of an investigation looks like more often than anyone admits.
The first two issues in this series argued that the lean, fair file protects both sides, and that when your own process falters, the strong move is to correct it in the open. Both of those pieces quietly assumed the process eventually arrives somewhere. This one is about when it doesn't — because in my experience a large share of workplace matters are credibility contests without corroboration, and the civil standard is honest about what that means. The Supreme Court settled the question in F.H. v. McDougall (2008): there is one standard, the balance of probabilities, and the evidence must be sufficiently clear, convincing and cogent to meet it. When it isn't, the truthful answer is that nothing has been proven — which is a statement about the evidence, and only about the evidence. The whole craft of the no-finding lies in writing it so that it stays that way.
Because there are two dishonest ways to end an unresolved matter, and each one gaslights a different party. Write "no wrongdoing occurred" and you have told the complainant their account was false — a conclusion the evidence never reached. Write in a register of lingering suspicion — the allegation could not be confirmed, in a tone that adds this time — and you have converted "not proven" into a sentence the respondent serves indefinitely. In my experience the memo has three jobs, and it is safest to give each its own sentence: state what was alleged; state precisely where the evidence ran out, with the standard named; and state what the outcome is not — not a finding that the complaint was false, and not a finding of misconduct. Some files support one further distinction, worth making explicitly whenever it is true: unsubstantiated — the evidence was insufficient — is not unfounded — the evidence affirmatively showed it did not happen. Knowing which of the two you are writing, and saying so, is the difference between an ending and a wound.
Then there is the harder version of this file, the one nobody likes to name. Sometimes the air in a matter carries a question about the process itself — a complaint that arrives the same week as a difficult performance conversation, a grievance whose timing does more talking than its contents. The two easy roads are both wrong. Pretend the question isn't there, and the file is naive; the respondent's claim of unfairness acquires real legs, because everyone in the room knew and the record didn't. Chase it openly, as an accusation, and you have manufactured a second complaint out of the first — and told a possibly genuine complainant that coming forward made them a suspect. The road between is the same discipline this series keeps returning to: treat tactical use as one hypothesis among others, and test it the way you test everything — timing examined, corroboration sought, motive weighed in both directions — then record, in one neutral sentence, that the question was asked and answered. The timing and context of the complaint were considered; the evidence did not support a conclusion that it was brought for a collateral purpose. I have come to believe that sentence protects the complainant more than anyone, because the question will be raised by someone — the respondent, the union, eventually an arbitrator — and a genuine complaint's best armour is a file showing it was already put, tested, and laid down in the same even voice as everything else. And on the rare occasion the evidence points the other way, the same discipline holds: it becomes a finding to be reasoned, never a label to be thrown.
Underneath all of this sits a structural unfairness I have been slower to see. In a unionized workplace the fairness machinery is built around the person in jeopardy — rightly, as far as it goes. The respondent arrives with representation, with notice, with someone at the table whose entire role is their protection. The complainant, who set the process in motion, very often walks in alone; they are not formally "facing" anything, so the apparatus does not attach to them, and sometimes the union representing the respondent is the same union they belong to. The file feels this imbalance even when nobody names it. If every procedural care in the record flows one way — notice to the respondent, representation for the respondent, confirmations to the respondent — the record itself tells a story about whose fairness the organization was tending. So the discipline is to document the complainant's side of the process with the same rigor the first issue gave to procedural facts: support offered, options explained, updates given, the outcome and what it does and does not mean walked through carefully. Not as courtesy. As part of the twenty percent that carries the weight.
The last issue ended with a question I could not resolve — the complainant who absorbs part of the cost when the process stumbles — and the first ended with another: the person left standing under a suspicion neither confirmed nor cleared. I notice now that they are the same question. When a matter ends without an ending, someone has to carry what the evidence could not settle, and the only decency left within the file's power is precision — to say exactly what was found, exactly what was not, and to refuse to let either party's dignity be spent covering the difference. That is not nothing. But I want to be honest about its limits, because the day after the memo, the two people in it still share a hallway. The file closes; the hallway does not. What happens between two people after the record ends is not a documentation problem at all, and I am coming to think it may be the more important half of this work. That is where this space is headed next.
If you have ever had to deliver a finding that found nothing — how did each side hear it? Reply and tell me. I read everything that comes back.
Sources & Further Reading
F.H. v. McDougall, 2008 SCC 53, [2008] 3 S.C.R. 41. The Supreme Court of Canada's confirmation that there is a single civil standard of proof — the balance of probabilities — and that evidence must be sufficiently clear, convincing and cogent to meet it. CanLII
Steven E. Abraham & Paula B. Voos, "Procedural and Distributive Justice in Sexual Harassment Arbitrations: Evolution of Decisions in the Union Context," Advances in Industrial and Labor Relations, Vol. 26 (Emerald, 2021). Emerald Insight