Why facts win cases: what a fact is, facts against inferences, assumptions and opinions, and the investigation cycle

Take an ordinary dispute: a bakery has a leak after a roofer repaired its roof. The rule of law is familiar. A person who does work for payment must do it with reasonable care, and must pass on warnings that the customer would want. Nobody in the case will argue about that rule. The argument is about whether the roofer warned the bakery that the gutter was rotten, when, in what words, and whether the water came in through the part he repaired or through the part he warned about. Those are questions of fact, and the lawyer who has the better grip on them is much more likely to win.

There are three reasons why facts carry so much weight. First, the facts decide which rule applies and how. The same rule gives opposite results on slightly different facts, which is why you saw in the Reading and Briefing Cases course that a holding is tied to its material facts. Second, in most systems the first court decides the facts and later courts rarely disturb them. An appeal court will correct a mistake of law, but it will seldom replace the trial judge’s view of who was telling the truth. A case that is lost on the facts at first instance is usually lost. Third, facts are expensive to rebuild. Evidence disappears: documents are deleted, memories fade, people move away. The time to get the facts is early.

A lawyer’s relationship with the facts therefore has two sides. One is discovery: finding out what happened, whether or not it helps your client. The other is proof: gathering the evidence that will show it to a decision-maker who was not there. A client’s account is the starting point and never the finish. People report events as they remember them, shaped by what they feel and what they hope. A careful lawyer takes the account seriously, tests it against everything else and is not surprised when parts of it do not survive.

In everyday speech, anything stated confidently is called a fact. In legal work you need a stricter test. A fact, for the purpose of investigation, is something that happened or exists and that could in principle be observed, recorded or checked: a person said these words, a document has this date, the stain is 40 centimetres across. A good test is: could two honest people with the same access agree on it?

None of these is bad in itself. Inferences are how we get from the evidence to the conclusion: nearly every legal case depends on them, because direct proof of what a person knew or meant is rare. Opinions from qualified experts can be evidence. Reports can lead you to the person who saw the event. The danger is mixing them up. When ‘the roofer ignored the gutter’ (an inference about his state of mind) is written as if it were a fact, the file starts to rest on something that nobody observed, and the next person to read it takes it for granted.

The habit to build is to go back to the fact under every conclusion. When a client says ‘the roofer was careless’, ask what he did. When a witness says ‘he must have known’, ask what he saw or heard that makes the witness say so. When a note says ‘it always leaks there’, ask how many times, and when. Then write down what was done or said, with its source, and keep your own inference separate and labelled.

A last distinction: first-hand and second-hand knowledge. A witness who saw the water on the floor has first-hand knowledge. A witness who was told about it has second-hand knowledge: what they can truthfully say is that they were told. The law of evidence in many systems restricts reports of this kind, in different ways. For investigation, the rule is simple: record who knows each fact first-hand, and go to that person.

Not every fact matters. A material fact is one that bears on a legal element or on the credibility of a witness, so that the outcome could differ if it were otherwise. The colour of the roofer’s van is not material. Whether he told the baker about the gutter is. Material facts are found by looking at the legal elements of the claim or defence: each one the party must prove is a place where facts are needed. Module 7 builds this into an evidence map.

A fact is disputed when the parties disagree about it and undisputed when they agree or when it is plain from documents. Time spent on undisputed facts is mostly time spent writing them down. The investigation’s real work is on the disputed material facts: the ones that matter and on which the sides differ.

Burden and standard of proof in outline. The burden says which party must prove a point; if the evidence is evenly balanced, that party loses on that point. In many systems the party who asserts a claim bears the burden of it, with exceptions. The standard says how convincing the proof must be. In civil disputes the usual standard is that the point is more likely than not (the ‘balance of probabilities’ or ‘preponderance’), and in criminal cases the prosecution must prove guilt to a much higher level, usually described as beyond reasonable doubt. Some civil matters, such as allegations of fraud in some systems, are treated more strictly in practice. These are the general patterns; the rules of your own system and court control.

Why does it matter to an investigator? Because it tells you what you must find. If your client bears the burden on an element, you need evidence for it, not just the absence of evidence against. If the other side bears it, you may need only to show that their evidence does not reach the standard, and your investigation can focus on testing theirs.

Fact investigation is not a straight line. It is a loop in which each step may send you back to an earlier one. A useful picture has seven steps.

1. Take the account. Listen to the client or the first source. Record the story in their words, and note what they know first-hand.
2. Identify the elements. Work out what must be proved (module 7). This tells you which facts matter.
3. List the sources. For each fact, where could evidence of it come from: documents, objects, digital records, witnesses, experts (module 2)?
4. Preserve. Secure the evidence before it is lost: send a preservation request, collect originals, photograph (module 2).
5. Collect and test. Interview witnesses (module 3), obtain documents and compare them with each account. Look for what supports and what contradicts (modules 4 and 5).
6. Organise. Put the material into a chronology, a cast of characters and an evidence map (modules 4 and 7).
7. Decide. Form a theory of the case, find the gaps, decide what to do next and when to stop (module 7).

The cycle explains why a good investigator keeps asking the same small questions at every stage: who knows this, how do they know, when did it happen, where is the record, what else would be true if it happened this way, and what would show I am wrong. The last question matters most. Investigators who look only for support get a one-sided file, which is no use when the other side does what you did not.

The Fact-or-Inference Sorter, in the lab below, is a first tool for step 1. Paste a draft account and it marks the sentences that read as judgements, inferences, assumptions or reports from others, so that you go back to the fact under each. It reads words, not truth, and it is no substitute for questions put to the source.