How do we really know what's true?

October 22, 2021

No matter how much we may disagree with each other, or how extreme our differences are, the one thing that we all have in common is that we all think that we’re right.

We believe that we’re the ones being rational, and that “the science” supports our claim, and that it’s others that are being irrational or biased.

It’s clear, then, that whatever being correct is, it’s more than seeming correct. It’s also more than going through what people consider as truth-finding rituals, like “looking at evidence” and “listening to those with experience or expertise.”

I don’t think people really consider this very much, but there’s a problem here: how can we be sure that our own beliefs don’t just have the illusion of truth, in the same way we imagine others’ have?

In this article I will take you through a few mainstream ideas of where truth comes from, and hopefully I will convince you that this not just a topic for philosophers and scientists, but something that is important to all of us in almost every part of our lives.

Logic

Our first stop is logic, which is a word often used but also somewhat misunderstood.

“Logic” usually refers to deductive logic, which is “the process of reasoning from one or more statements (premises) to reach a logical conclusion.” In deductive logic, we combine statements together to create new ones, and use rules to determine our new statements are true or not.

Here is a classic example of a simple deduction:

  • All humans are mortal. (First premise)
  • Socrates is a human. (Second premise)
  • Therefore, Socrates is mortal. (Conclusion)

The first two lines are premises, introduced into our argument by simply stating them. Then, we can combine them together to create a new statement, which is our conclusion.

Deductive logic is a very strong system. A conclusion in logic is guaranteed to be true if:

  • its premises are true; and
  • its argument is valid, i.e. free of mistakes

Unfortunately, Socrates cannot escape his mortality!

Despite its strength, many people overestimate the actual applicability of logic in practical situations, for three main reasons:

It only kicks truth down the road. The first requirement for a correct conclusion is correct premises. But those premises might themselves be conclusions of other arguments, and so on… Logic is good for combining statements, but in many cases we need a different kind of knowledge to start from.

Invalid arguments can be hard to detect. Humans are fallible, the world is messy, and our brains are optimised for fast jumps to conclusions. In the real world, often the core logic of an argument is obscured under layers of rhetoric, or terms with unclear definitions.

The absence of true is unknown, not false. A correct argument gives a true conclusion, but an incorrect argument does not give a false one. Without more information, we just don’t know either way yet.

This mistake is made often. Let’s look at this:

  • If Jen is a ski instructor, then she has a job.
  • Jen is a ski instructor.
  • Therefore, Jen has a job.

This argument is definitely valid. But what if the premises are false, and Jen isn’t a ski instructor? We can’t just take the conclusion and turn it around:

  • If Jen is a ski instructor, then she has a job.
  • Jen is not a ski instructor.
  • Therefore, Jen has no job. (Invalid)

This is invalid; we don’t know if Jen has a job or not. Deductions work just like “if … then …” statements: they only give information when the requirements (true premises & valid logic) are satisfied. If not, we can’t say anything about the conclusion.

We can write this form generally, where P and Q stand for any statements:

  • If P, then Q.
  • Not P.
  • Therefore, not Q. (Invalid)

These invalid argument forms are called fallacies, and this is the reason that identifying fallacies isn’t important for disproving things, since fallacies aren’t signs saying “wrong;” more “further work needed.”

Note that, of course, we can prove that things are false in logic, but we do it by showing that the opposite of the claim is true.

Scientific method

So, logic is powerful, but not a complete system of truth. When we turn our focus to the real world, we need a different kind of truth-finding system. This is where empirical observation and science comes in.

Largely stereotyped as an exclusive and technical practice, science is actually something very fundamental—something that we all do, every day of our lives: describing and predicting reality with general rules.

Scientific practice is codified in a general way into “the scientific method,” which looks roughly like this:

  1. Make careful observations. Attempt to measure accurately, and use tools or techniques with low bias or error.
  2. Construct a model or “hypothesis” which explains the observations.
  3. Continually refine, or discard, the model based on new observations.

The idea of science wasn’t handed to humanity. In fact, the modern concept of it has only existed for a few hundred years. Most people learn it roughly in school, but there are some important things to point out.

The first is that the scientific method is a cycle. It’s not possible us to be absolutely certain that we’ve fully figured something out (though we can know generally how powerful theories are). New observations reduce the number of hypotheses available to us, because hypotheses must be able to explain all known observations. Over time we can eliminate more and more explanations and theories as false.

Secondly, there is the jump between the first two steps: making observations to constructing a hypothesis. There aren’t strict rules about how to find or form a hypothesis (it is one of the many uses of creative thinking in the sciences). There also may be many, many different possible explanations for the observations seen so far.

In the past, it was widely thought that the step from observation to hypothesis possessed some logical quality—that it was a kind of “inductive reasoning” which conveyed an amount of truth, and this is what “powered” the truth of science. Indeed, most people today probably intuitively believe induction and generalisation to be a valid way to discover truth.

Unfortunately, this concept was dealt some critical blows over the last few centuries, to the point where many scientists today consider it to be entirely meaningless.

Basis of science

In the 18th century, Scottish philosopher David Hume concluded that all scientific laws, extending to the very concepts of cause and effect, were entirely logically invalid. He argued that if we repeatedly observe that some event B follows event A, there is no way we can logically conclude that “A causes B,” or “B will always follow A.” Any attempt to do so will depend on the idea that things that happened in the past will continue to happen in the future, which is not something that can be proven. We can’t even prove that “B is likely to follow A,” since that also depends on the probabilities staying the same.

As APXHARD puts it:

If I believe that there are laws of physics which govern the evolution of the material world, there is not a single experiment which can prove this. … It is only faith in something beyond physics which makes it feasible for a person to believe that the laws of physics don’t change in the shadow of Jupiter, or that they won’t change tomorrow. What experimental evidence rules out the idea that a week from now, physical laws will change?

This presented a major philosophical problem with science, and lots of thinkers of the time tried to solve the “problem of induction” and find a basis for science. The resolution, popularised by Karl Popper in the 20th century, is not to try to bend the spoon, but instead to realise the truth: there is no spoon. Induction, he says, doesn’t exist as a logical principle; it is an illusion, and does not ‘power’ science. The true power of science comes from the continual attempts to refute hypotheses by discovering things which contradict with them.

  • If my hypothesis is true, then I should observe X.
  • I observed not X.
  • Therefore, my hypothesis is false.

This is logically valid, and grounds science on the much more stable foundations of deduction, instead of induction. It is called the hypothetico-deductive model of the scientific method.

Notice though, that this grounding is backwards to how we think of logical proof. As I said above, we can never strictly be absolutely certain about a scientific law is correct. In logic, a correct argument gives ‘true’, and an incorrect one gives ‘maybe’. In science, the situation is different. To quote Einstein:

The scientific theorist is not to be envied. For Nature, or more precisely experiment, is an inexorable and not very friendly judge of his work. It never says ‘Yes’ to a theory. In the most favorable cases it says ‘Maybe,’ and in the great majority of cases simply ‘No.’ … Probably every theory will some day experience its ‘No.’ Most theories, soon after conception.
–Albert Einstein

We might think of logic as a {true, maybe} system and science as a {maybe, false} system.

A great way to bring this mindset into your life is to examine beliefs under the question: what would it take to convince me, or someone else, otherwise? What evidence would I need to see, from what sources? How likely do I think that evidence exists, but I just haven’t seen it? Seeking out refutation is how understanding evolves, so if you can’t think of anything even in theory that would change your mind, there’s a good chance you aren’t thinking scientifically.

Conclusion and summary

In addition to the conceptual difficulties with finding truth that we’ve discussed, there are also many practical ones. Experimental design, interpretation of data, and conflicts of interest can pose great challenges to researchers. Veritasium’s video Is Most Published Research False? gives a great overview of some of these problems.

I hope I have managed to convince you that being correct can be extremely difficult. It takes a lot of care to cut away explanations and converge on the truth.

But I hope I have also convinced you that there are ways of thinking and working which get us closer, and that these ideas can be applied to the reasoning that we do every day:

  • there are two kinds of claims, logical and scientific
  • logical claims are built from other true claims by valid arguments
  • scientific claims are made by carefully making hypotheses and then testing them
  • scientific claims can never be strictly proven true, but can be very powerful
  • we make scientific claims more powerful with better measurement and conflicting evidence

And finally, through the skepticism I’ve encouraged in this article, we should also be optimistic and remember how far we’ve come. In the last few centuries (before which almost no science existed), our understanding of the world has accelerated an absolutely incredible amount. Thanks to a philosophical mindset shift towards measurement and reasoning, we now have many highly powerful theories which explain the world to a very high degree of accuracy. So while truth may be elusive, with the right attitudes we can get closer than we could have ever dreamed.

Further reading

← All posts