what is a theory?

Recently, Michelle Bachman said: “Why would we forestall any particular theory? Because I don’t think that even evolutionists, by and large, would say that this is proven fact. They say that this is a theory…” This is really just an echo of the continuous drone of the “evolution is just a theory” mantra that has been coming from conservatives from time-out-of-mind. What is worse, though, is that I still hear liberals responding by saying, “Evolution isn’t a theory! It’s a fact!” This tells me that it’s time to take a step back and talk about what the word “theory” actually means.

First, let’s take a look at the relationship between three ideas that are used a lot in science: a theory, a hypothesis, and data.

Start at the top.  In science, “data” is a measurement of a relationship between things.  It’s always a relationship.  A simple measurement like “Bob is 6′ tall” is not data… or at least, it’s not the kind of data that is useful in science.  Data in science is always a measurement that can be used to come up with a statement about a relationship between two things.  The average man is taller than the average woman.  The average lifespan of people is increasing over time.  The number of people who die of gangrene is decreasing over time. And so on. These are statements about actual measurements of relationships in the world. Because they are numerical measurements, they can be measured over and over again by different people, and if the measurements are done right, then the results will be the same every time. This is why scientists like measurements: they can be verified. If you don’t trust a result that someone claims to be true, then you can do the measurement yourself.

The word “fact” doesn’t appear in the graph above, but if there is any idea on that graph that is closest to the idea of a “fact” it is the data. The data is objective, and it is either right or wrong. You can get into arguments about whether the measurement was done properly, or whether the fact measured is the “whole picture” or not. But data is not an opinion, and it is not an explanation of anything: it’s a statement about something that has been measured in the world.

A hypothesis is a statement that predicts a certain relationship; that is, it predicts a certain kind of data.  A hypothesis is an expectation of the way things are.  The statement “men are taller than women, on average” is a hypothesis. You can then go out and take measurements and they will either confirm the hypothesis, or disconfirm the hypothesis. In other words: the data will either agree with the prediction or disagree with the prediction.

A theory is an explanation of why things are the way they are. Notice that a hypothesis (e.g. “Men are taller than women”) is not an explanation. It’s a statement about a measurement that you expect to be true…. but it tells you nothing about why you might expect it to be true.  The theory is the explanation of why.  It is the set of ideas about how things happen or how they are structured “out there in the world” to produce the results that we see.

So, suppose you start with a theory (e.g. “men evolved to be hunters and protectors”). That theory, by itself, cannot be tested.  Let me repeat that:

THEORIES ARE NEVER DIRECTLY TESTED.

You have to first use the theory to generate a hypothesis. So, you use some kind of reasoning about what sorts of observations would logically result if your theory were true (e.g. “if men evolved to protect, then they should be stronger and faster and bigger”).  So the theory generates a hypothesis.  That hypothesis is an expectation about the way you expect a measurement to turn out (e.g. “Men are bigger than women.”). Then, you can go out and look at the data: the measurements of the relationship as it actually is in the world.  This data will either agree or not agree with the hypothesis. If the data disagrees with the hypothesis, then it is said to “disprove” the theory.

This is the classic way of understanding the “theory” of how science works, in an ideal scenario.

Notice something very important about the graph above, though: the word “proves” never appears. Why?  Because you may show that one of the hypotheses produced by the theory is true, but there always could be some other hypothesis that the theory produces that could end up being false.  And you can never know for sure that you’ve tested all of the hypotheses that a theory can produce. That’s why scientists say: “A theory can never be proved, only disproved.”

The simplest way of understanding this is to think about it this way: Suppose I tell you “every human being who will ever be born can click his tongue.”  If you go through 10 people and all of them can click their tongues, have you proven me correct? No, because the 11th person might not be able to. If you go through 1 million people and all of them can click their tongues, have you proven me correct? No, because the next guy might not be able to.  Every additional person who can click his tongue adds support for the hypothesis.  But the moment you come across just one person who cannot, you have proven me wrong.  That’s how science works: it only takes one bad experiment to show that a theory is wrong, but nothing is ever enough to demonstrate that the theory is infallible.

Now: that’s the standard theory of how science works. But let’s look at how science works in the real world for a moment.

In reality, complex theories always have a lot of ambiguity. You can come up with a general theory that outlines how things work, but then realize that when it comes right down to it even that theory has a lot of possible ways that the details can pan out. For example, the theory of evolution as put forth by Darwin simply said that there is some mechanism by which traits are passed from one generation to the next. But he didn’t know about DNA. So at the time that he produced the theory, there were a huge number of possible specific theories of how traits might be passed from one generation to the next, all of which were consistent with the general theory of evolution that Darwin outlined.

Also, think about the relationship between a theory and a hypothesis: you try to come up with a specific prediction about the way that the data should look, given what you know about the theory. This is a pretty mushy kind of relationship. What if your initial intuition makes you think that the theory would produce one hypothesis, and then when you think about it more you realize that it could produce a different hypothesis? With really complex theories, this kind of thing happens all of the time. For example, initially you might (reasonably) think that since mutation rates are low and happen continuously, evolution over time will always be gradual and continuous. But as it turns out, that isn’t necessarily a prediction of the theory of evolution. As people began to think more critically about the theory, they realized that there are a number of circumstances where Darwin’s theory is completely consistent with a predictions of rapid jumps, as well as periods of consistency, in the evolutionary process.  So over time, scientists can actually learn about what hypotheses a theory is really able to produce.

 

This graph, above, is more like what science looks like in the real world.  People come up with a general theory that describes some basic, underlying principles about how they think the world works.  Evolution is this kind of theory: a general statement about some bedrock underlying principles of how traits are passed on and change over time.

But what about the details? A general theory can often “fill in the details” in a number of different ways.  Usually a scientist will pick one way and use that to generate hypotheses that can be tested with data.  Then, if one of those hypotheses is tested and the data comes back disconfirming the hypothesis, it disproves the specific theory, but it does not necessarily disprove the general theory. The question then becomes: is there a different specific theory that can be produced that is consistent with the new data, but that still is based on all of the underlying assumptions of the general theory? If the answer is yes, the general theory can move forward, with new insights gained by what we have learned about some of the details in the specific theory.

This is how real science happens all the time. This is how theories evolve, and are adjusted over time. Sometimes, data comes along that is absolutely impossible to account for with any specific theory that can be generated from the general theory… and then the general theory is abandoned.  But as they say, “the devil is in the details,” and often the data acts more as a way of educating us about the details of a specific theory than causing us to throw out the general theory altogether.

So, to sum up:

  • No theory is ever proved.
  • Theories are never directly tested, they are only tested by first coming up with hypotheses that link the theory to data
  • General theories usually can give rise to specific theories, and disproving one specific theory doesn’t necessarily disprove the general theory on which it was based. Instead, it is just used to create a better specific theory from the same general principles.
  • Evolution is a general theory. It IS a theory, and it has NOT been “proved.” But it also has not been disproved. In fact, there is so much overwhelming evidence supporting the general theory that saying “evolution is just a theory” is like saying “the idea that the earth is round is just a theory.”  It IS a theory, but there is no “just” about it.

Not to be too partisan, but: we all expect conservatives to get this stuff wrong.  Let’s make sure we liberals get it right, too.

 



Leave a Reply

Your email address will not be published. Required fields are marked *

Comment

You may use these tags : <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>

Trending Articles

Keep up with my writing!You will only be notified about new articles. No ads, no petitions, no digests, no nonsense.