EVIDENCE

It is undesirable to believe a proposition when there is no ground whatsoever for supposing it true.

• Bertrand Russell

Evidence helps us prove what is likely to happen or likely to be true or false. It is a fact or collection of facts that provides support—affirmation, confirmation, corroboration, proof, substantiation, verification—for the truth or falsity of a claim. Evidence can take many forms:

• Direct sense evidence — using sight, sound, touch, taste, or smell to observe spatiotemporal entities directly. Example: visiting Paris and seeing the Eiffel Tower with your own eyes.
• Indirect sense evidence — observing through a reliably calibrated device or instrument, such as a magnifying glass, periscope, camera, video recorder, binoculars, microscope, telescope, or meter. Example: seeing photographs or videos of the Eiffel Tower without having been to France.
• Testimony of trusted others — relying on the accounts of people we find credible. Example: parents describe their trip to Paris and their visit to the Eiffel Tower, and we believe them.
• Expert testimony — the informed judgment of specialists in a domain. Example: Gustave Eiffel, the tower’s engineer, testifies to other engineers that the structure can support an elevator system.
• Authoritative explanations from science — well-established theories and laws that explain phenomena. Example: Newton’s law of gravity explains why the Eiffel Tower does not float away.
• Logical or mathematical entailment — conclusions that follow necessarily from accepted truths. Example: if the Eiffel Tower is in Paris, and Paris is in France, then the Eiffel Tower must be in France by the transitive property of “located in.”
• Arguments — structured reasoning can itself serve as evidence. For instance, a reductio ad absurdum shows that a claim leads to an absurdity, contradiction, or impossibility, thereby falsifying it.

The weight we give to evidence varies with its reliability. Direct sense evidence from sober, sane observers under good conditions typically provides strong support. Data gathered by researchers in physics, chemistry, or biology is rightly taken as true when the methods are sound. Other claims—philosophical, theoretical, or purely subjective—require forms of support that go beyond pointing to a spatiotemporal object, and some claims may only have evidence acceptable to the person experiencing them.

Whatever form evidence takes, the core question is always the same: What is the evidence that supports the truth of this claim?

The methods of science

There are in fact two things: science and opinion; the former begets knowledge, the latter ignorance.

• Hippocrates

"What experiment can I do to figure this out?"

In 1986, the space shuttle Challenger exploded on launch, killing all astronauts aboard. After the disaster, a commission was formed, and Richard Feynman demonstrated that the fuel booster rockets were unsafe in cold temperatures. The temperature at takeoff was 32°F. During a launch, vibrations cause the rocket joints to move. Inside those joints were rubber O-rings. Feynman performed a simple experiment: he squeezed a spare O-ring in a C-clamp and dipped it in a glass of ice water at 32°F, showing that the rubber did not expand. Without resilience at that temperature, the O-ring could not fill the gap in the expanding booster joints, leading to the explosion. The experiment was not complicated—it merely isolated the key variable and tested it.

This exemplifies the scientific approach: strive for objectivity. Scientists try to describe the world as it is, not as they wish it to be. They seek to answer “why” and “how” questions and predict natural phenomena using methods of scientific integrity. As Adam Smith wrote, “Science is the great antidote to the poison of enthusiasm and superstition.”

The scientific process includes the following steps—though trial and error, luck, and intuition also play a role:

• Problem or observation — We notice a phenomenon or a practical problem and wonder what happens and why.
• Guess why — We form a hypothesis: a possible explanation that can be proved or disproved by testing against experiment and observation. The guess must be measurable and consistent with established evidence.
• Predict consequences — We deduce the logical implications of the guess. If the guess is correct, what else must be true?
• Test — “If I do this, what will happen?” Testability is the key. We compare predicted consequences with experiment, evidence, and observation, repeating the test to guard against error, fraud, coincidence, or changed circumstances. We report results honestly. The more evidence that agrees with the guess, the more likely it is correct. If it disagrees with experiment, it is wrong. As Richard Feynman said:

It does not make any difference how beautiful your guess is. It does not make any difference how smart you are, who made the guess, or what his name is — if it disagrees with experiment it is wrong.

Darwin understood that observation without a guiding theory is aimless. On board the HMS Beagle, he wrote in a letter to Henry Fawcett:

About thirty years ago there was much talk that geologists ought only to observe and not to theorize; and I well remember someone saying that at this rate a man might as well go into a gravel pit and count the pebbles and describe the colours. How odd it is that anyone should not see that all observation must be for or against some view if it is to be of any service!

We don’t merely observe; we observe with purpose, guided by a hypothesis or a theory about what matters. Charles Munger extends this to information gathering:

... you have to have some idea of why you're looking for the information. Don't read annual reports the way Francis Bacon said you do science ... where you just collect endless [amounts of] data and then only later do you try to make sense of it. You have to start with some ideas of reality. And then you have to look to see whether what you're seeing fits in with that basic thought structure.

When multiple explanations make the same predictions, Occam’s Razor advises: “Entities should not be multiplied unnecessarily.” Prefer the explanation with the fewest unproven assumptions, until new evidence forces otherwise. As Einstein paraphrased it: “Theories should be as simple as possible, but no simpler.” Occam’s Razor does not rule out other explanations; it simply keeps us from unnecessary complexity while evidence is incomplete.

Finding evidence from the past

Study the past if you would divine the future.

• Confucius

John is thinking of hiring a new manager. Does the manager’s past record matter? Warren Buffett says:

The best judgment we can make about managerial competence does not depend on what people say, but simply what the record shows. At Berkshire Hathaway, when we buy a business we usually keep whoever has been running it, so we already have a batting average. Take the case of Mrs. B. who ran our Furniture Mart. Over a 50-year period, we'd seen her take $500 and turn it into a business that made $18 million pretax. So we knew she was competent ... Clearly, the lesson here is that the past record is the best single guide.

Then you run into the problem of the 14-year-old horse. Let's say you buy The Daily Racing Form and it shows that the horse won the Kentucky Derby as a four-year-old. Based on past performance, you know this was one hell of a horse. But now he's 14 and can barely move. So you have to ask yourself, "Is there anything about the past record that makes it a poor guideline as a forecaster of the future?"

To decide whether past evidence is representative for the future, ask:

• Observation: Will past or present behavior continue? How long can it continue?
• Explanation: Why did it happen? What are the key factors? Start with a hypothesis; compare its implied consequences with appropriate evidence for and against. You must understand the equation—the key variables and their relationships.
• Predictability: How predictive is the past evidence? Are the conditions now and in the future likely to change? Ensure the evidence isn’t the result of mere chance. What worked before could be luck.
• Continuation and Change: What is required for the past record to continue or for the goal to be achieved? What must happen? What must not happen? What forces could alter the outcome—and how likely are they? What antidotes exist for what we don’t want to happen?
• Certainty and Consequences: How certain am I? What single event am I betting on that must—or must not—happen? What are the consequences of being wrong?

Falsify and disprove

All our commonsense ideas should always be open to criticism.

• Karl Popper

Scientific results always carry some probability. Tomorrow may bring new evidence. Instead of only trying to verify a statement, it is often better to try to prove it false. A single confirming instance supports a statement but does not prove it; a single contradictory instance refutes it. Einstein said: “No number of experiments can prove me right; a single experiment can prove me wrong.”

Consider the statement: “All swans are white.” To test it, go looking for non-white swans. If we find one black swan, the statement is disproved. The more white swans we find, the more supported the statement becomes—but it is never proven beyond all doubt.

Or take: “The universe is no more than 10,000 years old.” We can falsify this by observing stars millions of light-years away; we are seeing them as they were millions of years ago. The statement fails.

In medicine, a claim like “The treatment worked—I’m cured” must face the question “Compared to what?” To reduce error and bias, randomized, double-blind, placebo-controlled studies are used. Subjects are randomly assigned to treatment or placebo groups, matched for age, physical status, and other factors. Neither subjects nor researchers know who receives the treatment, and then results are compared. This design isolates the true effect of the treatment from suggestion, salesmanship, or imagination.

A claim of absolute safety—“There is no risk in using this medicine”—is not the same as evidence of safety. The absence of evidence of harm is not evidence of absence. It only means the substance appears safe based on current knowledge, just as an archaeologist’s failure to find a fossil doesn’t prove it doesn’t exist.

Charles Darwin made a habit of looking for evidence against his own ideas:

I had, also, during many years, followed a golden rule, namely, that whenever a published fact, a new observation or thought came across me, which was opposed to my general results, to make a memorandum of it without fail and at once: for I had found by experience that such facts and thoughts were far more apt to escape from the memory than favourable ones. Owing to this habit, very few objections were raised against my views which I had not at least noticed and attempted to answer ...

Falsification can also take the form of a reductio ad absurdum argument, which shows that a claim leads to an absurd or impossible consequence, thereby proving it false. For example:

1. If it were true that Noah fit two of every species of living thing on the ark, then the ark would need to have been the size of Australia—an absurdity inconsistent with the biblical description (roughly 520 feet long by 86 feet wide by 52 feet high).
2. We cannot accept that the ark was the size of Australia.
3. Therefore, it is false that Noah fit two of every species on the ark.

When many pieces of evidence agree with our explanation, we must still ask: might the same evidence support other explanations? Look for evidence that disproves your explanation specifically. Don’t waste time on ideas already disproved or that cannot be disproved. Always ask: “What test can disprove this?” For instance, “There is life on planet Zeta” cannot be tested with current technology. That doesn’t mean it is false—only that we have no way to evaluate the claim today.

Theories grounded in observations take priority over purely abstract theories, because observations can falsify them. Galileo’s telescopic observations of the moons of Jupiter helped overturn the Earth-centered model of the universe.

Engage in self-criticism. Question your assumptions. Explain the opposite of your beliefs. Ask: “If I’m wrong, how will I know? Why might an opposite theory be correct? Assuming my answer is correct, what evidence would cause me to change my mind?” Then actively seek that evidence. Because we often miss our own weaknesses, seek objective feedback from people who can help us improve. Measure your outcomes against your original expectations, and find mistakes early enough to correct them before harm occurs.

Charles Munger describes the habit of backward thinking as a way to force objectivity:

The mental habit of thinking backward forces objectivity — because one of the ways you think a thing through backward is you take your initial assumption and say, "Let's try and disprove it." That is not what most people do with their initial assumption. They try and confirm it. It's an automatic tendency in psychology — often called "first-conclusion bias". But it's only a tendency. You can train yourself away from the tendency to a substantial degree. You just constantly take your own assumptions and try and disprove them.