In an earlier discussion, I tried to understand what we mean when we say that something is true. The position I eventually arrived at was fairly simple:

Truth is correspondence with reality.

If I say, “There is a cup on the table,” the statement is true if there really is a cup on the table. But that simple definition creates another problem. Reality may exist independently of us, but our access to it is limited. We see only a part of it. We measure it through instruments. We remember imperfectly. We interpret what we observe. Sometimes we are mistaken. Sometimes information is simply unavailable. So even if truth depends on reality, how do we determine whether our claims actually correspond to reality? This is where evidence enters the picture.

Evidence Connects Our Claims to Reality

Suppose I wake up in the morning and say, “It rained last night.” How do I know? Perhaps I look outside and see that the road is wet. The wet road is an observation. And because a wet road is something we would expect after rain, that observation supports my claim. This gives us a useful starting point:

Evidence is information or observation that gives us a reason to increase or decrease our confidence in a claim.

This definition is slightly different from saying that evidence proves something. Evidence does not necessarily prove that a claim is true. It gives us a reason to believe the claim more strongly, or sometimes less strongly. The distinction is important. The road may be wet because it rained. But perhaps municipal workers washed it early in the morning. Perhaps a water pipe burst. Perhaps sprinklers were running. The wet road therefore supports the rain claim, but it does not establish it beyond doubt.

How Do We Know Whether Something Is Relevant?

Simply being connected to the same topic is not enough. Suppose someone claims, “This student understands statistics.” We may observe that:

• the student attends every lecture;

• the student owns several statistics books;

• the student scored highly on an examination;

• the student can explain a statistical concept;

• the student can apply the concept correctly to an unfamiliar problem.

All these observations are connected in some way to statistics. But they do not provide equally relevant evidence for understanding statistics. Attendance may make understanding more likely, but a student can attend every class without understanding much. Owning books tells us even less. Being able to explain a concept and apply it to an unfamiliar situation is much more directly connected to what we mean by understanding. So relevance requires something stronger than a topical connection. A useful test is:

Would knowing this information reasonably change my confidence in the claim?

If the answer is no, the information is probably irrelevant.

Ask:

Would I expect to observe this if the claim were true?

and then:

Would I also expect to observe it if the claim were false?

This second question is extremely important.

Evidence Should Help Us Distinguish Between Possibilities

Consider the wet road again. If it rained, a wet road is quite likely. But even if it did not rain, a wet road is still possible. So the observation helps us somewhat, but not enormously. Now suppose a properly functioning rain gauge recorded 25 millimeters of precipitation overnight. That observation would be much harder to explain if there had been no precipitation at all. The evidence therefore becomes stronger. This suggests a deeper way of thinking about relevance:

Relevant evidence helps us distinguish between a situation in which the claim is true and one in which it is false.

The better it does this, the more useful the evidence becomes. This also exposes a common mistake in reasoning. People often say:

“My claim can explain this observation, therefore the observation proves my claim.”

But that is not enough. The better question is:

Can some other explanation account for the same observation equally well?

Being Consistent With a Claim Is Not Enough

Imagine that someone says:

“There is an invisible creature in this room.”

You look around and see nothing. The person replies:

“Exactly. You cannot see it because it is invisible. That supports my claim.”

The absence of a visible creature is certainly consistent with the claim. But the same observation is also exactly what we would expect if there were no creature at all. So the observation does very little to distinguish between the two possibilities. This is why:

An observation being compatible with a claim does not automatically make it strong evidence for the claim.

Whenever evidence is presented, we should therefore ask:

What alternative explanations are possible?

The easier it is for alternative explanations to produce the same observation, the weaker the evidence generally becomes.

Evidence Has Strength

Once we accept that evidence can support or weaken claims, another question naturally appears:

Is all evidence equally strong?

Clearly not. Suppose we want to know whether Rahul was in a classroom at 10 AM. We may have the following information:

• Rahul’s bag was found in the classroom.

• A student says she saw Rahul there.

• Three independent students say they saw him.

• CCTV shows Rahul entering at 9:55 AM and leaving at 10:20 AM.

• His access card was used at the classroom entrance.

• His phone-location record places him in the same building.

Each of these may be relevant. But they do not deserve the same weight. Rahul may have left his bag there earlier. A witness may be mistaken. A CCTV recording may provide stronger evidence, assuming the recording, timestamp, and identification are reliable. And several genuinely independent pieces of evidence pointing in the same direction may strengthen the case further. So evidence is not merely:

present / absent

It can be:

weak / moderate / strong

depending on how much it should change our confidence in a claim.

What Makes Evidence Strong?

There is no single property that makes evidence strong. Several things matter.

1. Relevance: How directly does the evidence bear on the claim? A student’s ability to solve and explain unfamiliar problems is more directly relevant to understanding than simply attending class.

2. Reliability: Can we trust the observation, measurement, source, or method through which the evidence was obtained? A measurement from a properly functioning instrument deserves more confidence than one from an instrument known to malfunction.

3. Alternative Explanations: How easily could the same evidence arise if the claim were false? The more plausible alternative explanations remain, the weaker the evidence usually is.

4. Directness: How many inferential steps lie between the evidence and the claim? Some evidence bears directly on what we want to know. Other evidence only indirectly suggests it.

5. Independence: Do several pieces of evidence come from genuinely different sources? Five people independently observing an event are very different from five people repeating something they all heard from the same person.

6. Quality of Method: How was the evidence produced? Careful observation, controlled conditions, reliable measurement, and transparent procedures generally provide stronger grounds for confidence than casual or uncontrolled observation.

Together, these factors help us answer the most important question:

How much should this evidence change my confidence in the claim?

More Evidence Does Not Necessarily Mean Better Evidence

We often assume that ten pieces of evidence must be better than one. Not necessarily. Suppose ten newspapers report the same story. At first glance, it appears that we have ten independent sources. But then we discover that all ten reports came from the same original anonymous source. We do not really have ten independent pieces of evidence. We have one piece of evidence repeated ten times. Similarly, ten poorly designed studies do not automatically become strong evidence merely because there are ten of them. Quantity matters, but quality and independence matter too. One highly reliable observation may sometimes deserve more confidence than a large collection of weak observations.

Is There a Hierarchy of Evidence?

It is tempting to create a universal ladder— this kind of evidence is always stronger than that kind. But that would itself be too broad a claim. The strength of evidence depends on what we are trying to establish. Suppose the claim is:

“This medicine reduces blood pressure.”

A large, carefully controlled experiment may provide much stronger evidence than one patient’s personal experience. But suppose the claim is:

“This patient felt dizzy after taking the medicine.”

Now the patient’s own report becomes highly relevant.

Similarly, experiments are extremely useful for some scientific questions, but they may not be available for historical questions. If we want to know whether a particular historical event occurred, documents, archaeological evidence, independent accounts, photographs, records, and physical traces may be the best evidence available. So instead of memorizing one universal hierarchy, it may be better to ask:

What kind of evidence would most strongly distinguish between the possible answers to this particular question?

A hierarchy of evidence therefore makes sense within a particular type of claim, but not necessarily across all claims.

Evidence Can Also Weaken What We Believe

There is another side of evidence that we often overlook. Evidence is not only something we collect to support our claims. It can also make us reduce our confidence in them. Suppose I believe:

“It rained last night.”

A wet road increases my confidence. But then I learn that:

• no rainfall was recorded by nearby weather stations;

• CCTV shows municipal workers washing the road;

• surrounding rooftops are dry.

Now the evidence begins to work against my original claim.

The intellectually difficult part is that we often like evidence when it supports what we already believe and resist it when it does not. But if the purpose of evidence is to help our claims correspond more closely with reality, then evidence must be allowed to move our confidence in both directions. Otherwise, we are not really using evidence to understand reality. We are using selected observations to protect a belief.

Evidence Does Not Contain Truth

This may be the most important point. Evidence itself should not be confused with truth. Reality is what it is. A claim attempts to describe some part of that reality. Evidence gives us access (always limited) to information that helps us judge the claim. So perhaps the relationship can be imagined like this:

Reality → Observation and Information → Evidence relevant to a claim → Confidence in the claim

Evidence is therefore a bridge between our claims and reality. Sometimes that bridge is strong. Sometimes it is weak. Sometimes we later discover that part of it was unreliable. This is why reasonable people can change their minds when better evidence becomes available without implying that truth itself has changed. What changes is our access to it and therefore the confidence our claims deserve.

From Belief to Evidence

This brings us back to a problem discussed earlier while thinking about truth. People frequently move directly from:

“I believe this”

to:

“This is true.”

Evidence inserts an important step between them. Instead of saying:

I believe it → therefore it is true

we can ask:

• What exactly is my claim?

• What evidence supports it?

• Is the evidence actually relevant?

• How strong is it?

• What alternative explanations exist?

• Is there evidence against my claim?

• How much confidence does the total evidence justify?

This produces a very different relationship with our beliefs. We do not have to stop believing things. We simply stop treating the existence of a belief as evidence of its truth.

A Simple Framework

Our earlier discussion on truth produced the framework:

Claim → Evidence → Scope → Confidence

The discussion on evidence allows us to look more carefully inside the second step. For any claim, we can ask:

1. Claim: What exactly am I claiming?

2. Evidence: What observations or information bear on this claim?

3. Relevance: Why should this observation affect my confidence in the claim?

4. Alternatives: Could the same observation occur if my claim were false?

5. Strength: How reliable, direct, independent, and methodologically sound is the evidence?

6. Confidence: Considering all the evidence for and against the claim, how strongly should I believe it?

None of this guarantees that we will always reach the truth. That would demand more from human knowledge than it can provide.

But it gives us something valuable:

a disciplined way of trying to make our beliefs correspond with reality as closely as our available evidence allows.

And perhaps that is what evidence is ultimately for.

Key Takeaways

1. Truth concerns correspondence with reality; evidence helps us judge whether that correspondence is likely.

2. Evidence is information or observation that can increase or decrease our confidence in a claim.

3. Evidence is always evidence in relation to a particular claim.

4. Being connected to a topic does not automatically make information relevant evidence.

5. Relevant evidence should help distinguish between the claim being true and false.

6. Being consistent with a claim is not the same as strongly supporting it.

7. Alternative explanations matter when evaluating evidence.

8. Evidence varies in strength.

9. Strength depends on factors such as relevance, reliability, directness, independence, alternative explanations, and quality of method.

10. More evidence does not automatically mean stronger evidence.

11. Several independent lines of evidence are generally more valuable than repeated information from the same source.

12. There is no single universal hierarchy of evidence; appropriate evidence depends on the kind of claim being examined.

13. Evidence can support a belief, weaken it, or leave our confidence largely unchanged.

14. Evidence should be allowed to challenge our own beliefs, not merely support them.

15. Evidence does not guarantee truth. It helps determine how much confidence a claim deserves.

16. A useful question is not merely “What evidence supports my claim?” but also “What evidence would make me change my mind?”

Perhaps the simplest way to remember the entire discussion is this:

Do not ask only whether you have evidence. Ask what the evidence is evidence for, how strong it is, what else could explain it, and how much confidence it truly deserves.