lukeprog.com: where every page is easy to read
Self-Help Science Writing Personal Religion Travel Self-Publishing

Religion

Why Do We Trust Science?

What’s so special about science?

Most people trust science to give us more reliable knowledge about the real world than any other method.

  • The FDA won’t allow new drugs on the market until they have been scientifically tested for safety and effectiveness.
  • Many groups require therapists to use only empirically supported treatments.
  • Today, all of our tools and gadgets are designed and tested with scientific methods. We would not want to fly a plane or drive a car that hadn’t been.
  • When schools teach about the real world (biology, geology, astronomy, psychology), they teach scientific findings only.
  • Marketers always want to claim “scientifically proven results” for their products, whether or not scientific methods were involved.
  • etc.

So it seems we trust science a great deal. But why? What’s so special about science? And more importantly, should we trust science?

A silly question?

Some people think those are silly questions. Why do we trust science? They might reply, “Because it works, duh!” or “As opposed to what? Blind guesses?”

After all, most of our knowledge has come from science. The gods didn’t tell us that e=mc2, or how to build an airplane, or how to treat ulcers, or how continental drift works. Science did! Blind guessing couldn’t tell us that our universe is about 14 billion years old. Science did!

Not a silly question

But other people say we can’t trust science. After all, there have been lots of scientific hoaxes. And science has been wrong before.

I don’t think it’s a silly question. Very smart people have wrestled with this question in the past century, and they are still wrestling with it.

Well, what is science?

The word science can mean:

  • a set of methods for getting knowledge (the doing of science)
  • the knowledge acquired by those methods (the is of science)

We’re asking about the knowledge. The is of science. Why do we trust scientific knowledge?

The answer turns out to be those scientific methods, the doing of science. For hundreds of years, the doing of science has produced more reliable knowledge than any other method.

Why? Well, because scientists keep choosing methods that give reliable knowledge, and abandoning methods that don’t. Let’s look at some traits of good science:

  • Science requires evidence.
  • Science applies critical thinking to the evidence.
  • Science tries to falsify a theory.
  • Science tests novel predictions made by a theory.
  • Science performs controlled, repeated experiments.
  • Science prefers simple explanations.
  • Science is tentative and progressive.

Each of these traits helps us seperate reliable claims from unreliable claims. Let’s look at each one to see why.

Science requires evidence

Let’s see how relying on evidence helps us seperate reliable claims from unreliable claims. We want reliable knowledge, not just a bunch of bullshit claims like astrology or phrenology.

Science says: “If you’re going to make a claim, back it up!”

This puts a burden on anyone who makes a claim. If I claim that the Flying Spaghetti Monster exists, it’s not the duty of everyone else to prove that no Flying Spaghetti Monster exists anywhere in the universe. That would be silly. No, I must show reasons to think that the Flying Spaghetti Monster exists.

This is called the burden of proof. The burden of proof is on whoever makes a claim.

And I can’t just show any reasons to think the Flying Spaghetti Monster exists. “He exists because I say he exists” won’t cut it. Nor will “He exists because I believe he does.” Nor will “He exists because I feel him.”

People actually do make those claims. Billions of adults claim to have an invisible friend for which they offer no evidence. But I think we all agree you need to show evidence for such a claim.

Otherwise, any claim is just as valid as any other. All I would have to say is, “I believe and feel our universe was created by a math student in a higher dimension,” and there would be just as much reason to believe me as anyone else who makes a claim without evidence.

Science wants reliable knowledge. That’s why science demands evidence. Evidence helps us seperate truth from bullshit. If we didn’t require evidence to back up a claim, then literally anything goes, as long as someone claims it.

Science applies critical thinking to the evidence

But evidence is cheap. Evidence is everywhere. If we don’t use critical thinking, almost anything can be evidence for anything.

Let’s use another Flying Spaghetti Monster example. We know that as the earth’s population has grown, so too has humanity’s average height. Pastafarians say this is evidence for The Flying Spaghetti Monster.

How? Well, here’s their explanation. The Flying Spaghetti Monster reaches down with his invisible, noodly appendages to touch his beloved humans on the head. But he only has so many noodles. As the earth filled with more and more humans, he had to press down on the heads of each one of us a little less often. There just weren’t enough noodles to go around.

So because he was pressing on each of our heads less often, we started to grow a bit taller on average. Makes sense, right? So, the correlation between the earth’s population and our average height is evidence for the Flying Spaghetti Monster, right?

No. The Pastafarians did not apply critical thinking to their evidence.

In this case, their argument is non sequiter. A correlation between the earth’s population and our average height does not imply a SpagheDeity.

In fact, that’s a bizarre and complicated answer. Many other, simpler explanations are available. Even if we couldn’t come up with a better explanation, the Flying Spaghetti Monster’s existence does not follow logically from a correlation between population and average height.

A real life example

Let’s try a real-life example. Christians claim that miracle healings performed in the name of God are evidence for God. After all, they prayed for God’s healing, and a person got better. That’s evidence for God, right?

There are a few problems with this idea. First, did they count their prayers? If they did, they would find that the vast majority are not answered. People recover from sickness at the same rate whether they are prayed for or not.

Second: If Christian miracles are evidence for the Christian God, then Muslim miracles must also be evidence for Allah, witch-doctor miracles must be evidence for animism, etc. Every religion has their own “miracles.” They can’t all be right, but they each make exactly the same claim, with the same evidence.

With a little critical thinking, we can see that miracle healings are not evidence for any particular gods.

Christians might say, “Well, but a miracle is evidence for some kind of supernatural force, right? How else could it have happened?”

Our ignorance is never evidence. Think about it. In the Middle Ages, we didn’t understand how lightning worked. Was that evidence for Zeus? Just because something looks mysterious or magical to us doesn’t mean it’s evidence of supernatural agents.

Ignorance means we don’t have knowledge. It means we shouldn’t make claims. Ignorance is the opposite of knowledge. If you’re ignorant, that is the time you shouldn’t make a knowledge claim.

So there’s plenty of evidence around. You have to apply critical thinking to find out what it’s evidence for. Critical thinking helps make knowledge more reliable.

Science tries to falsify a theory

If we’re going to test whether or not a theory is true, it should be falsifiable. To see why, let’s look at three theories that are not falsifiable:

  • Theory 1

    The Flying Spaghetti Monster created the universe at noon yesterday. He created all of us with “lifelong” memories that match all kinds of details in the world: that scratch on your knee, the shape of the lamp in the hall.

    He created light in transit from distant stars to look like they’d been burning for billions of years. He created millions of books that seem to have been written long ago, describing a detailed history of the world and its people. He even planted fossils and city ruins and time capsules in the ground as if they’d come from the distant past. He instantly decayed isotopes in the rocks and fossils so that if we try to date them it will seem they are thousands or millions of years old.

  • Theory 2

    An all-powerful, invisible, eternal being named Roger set our universe into motion 14 billion years ago. He let physics and evolution take their time to produce us humans. He let cultures and religions evolve.

    Every now and then, he messes with us. He causes someone to suddenly get better or suddenly die, but not often and not with any consistency. He doesn’t respond to requests. He doesn’t let himself be known. He just likes messing with us and watching what we do. He’s very amused by how many gods man has invented in his own image.

  • Theory 3

    There are an infinite number of universes, each one with slightly different values for things like the strength of gravity or the weight of a proton. We happen to be in one that had the right values to—after 14 billion years—produce self-aware life forms like ourselves. All other universe will always be invisible to us.

I could write these all day. And maybe you see the problem I’m getting at.

We can’t falsify any of these, because they could fit any evidence we come across. There are an infinite number of theories that can fit all the evidence we have. We need to be able to test a theory for it to be useful.This isn’t quite true. A couple theories, called “properly basic beliefs,” can’t be tested. For example, consider the statement “I exist.” What evidence could falsify that claim? Or: “There is an external world.” How could we conceivably falsify that? So, a few basic beliefs about the world’s existence are needed to ask any questions about the world at all. But that doesn’t mean we should assume anything else without question, like the existence of exotic gods.

Good theories

Gravity is a good theory this way. How can we test gravity? Drop a bowling ball. Does it fall to the ground? If not, there might be a problem with our theory of gravity. We can easily test our theory of gravity in a million different ways, because the theory makes very specific predictions about how things should happen if the theory is true.

The theory of common descent—related to the theory of evolution—is another good theory this way. It says that all life evolved from a common ancestor.

For example, millions of years ago there was a primitive kind of ape. Some of them got isolated and evolved to fit their environment. After millions of years, they became chimpanzees. Others got isolated elsewhere and evolved to fit their environment and became humans.

How can we test this? Well, the theory says we should find transitional fossils between the ancient ancestor and modern humans. Some fossils should be more like apes than humans, but not quite like either. Other fossils should be more like humans than apes, but not quite like modern humans.

We’ve found dozens of such transitional fossils—a whole spectrum of development. The theory made a prediction, we tested it, and it turned out to be true.

We also look for ways to falsify this theory. If we found a human-like fossil in rock that was older than all mammals, that would falsify the theory. We’ve dug up thousands of such rocks. Guess what? No human fossils. If we find something like that, we’ll have to come up with a new theory, or at least seriously change the one we’ve got.

Bad theories

Bad theories offer no way to disprove them. They will fit any evidence we find. As we saw before, this means they’re no better than the infinite number of theories that can fit all the evidence equally well.

I grew up being taught a bad theory like this. It’s called Young Earth Creationism. Here’s what I was taught:

  • God created the universe about 6,000 years ago.
  • He created all animals and humans in more or less their present form.
  • About 5,000 years ago, he helped a guy called Noah put 2 of every species on a boat so they would survive as he flooded the earth to kill all the evil people.
  • How did light from stars 5 billion light-years away get to Earth in less than 6,000 years? When God created the stars, he also created their light all the way between them and the earth so we could see them right away.
  • What about fossils we find that are millions of years old? Well, God created them to look like they were that old, because he doesn’t want to force us to believe in him through the overwhelming evidence. He wants faith instead.
  • But wait, all those animals couldn’t fit on Noah’s boat! The Bible gives the dimensions for the boat, and it’s simply not possible! Well, God made a miracle happen, and they all fit. You know, like how he made Jonah fit in the belly of a fish and not be dissolved by its stomach acid.

I was actually taught this in elementary school! All this was in a “physical sciences” book published by Bob Jones University Press.

See the problem with the theory of Young Earth Creationism? The problem is that no evidence can falsify it. The answer to all contradicting evidence is: “Magic. God did it with his invisible magic.”

That makes Young Earth Creationism exactly as likely as an infinite number of other theories that can’t be falsified. Which is to say, not likely at all. At all.

Science tests novel predictions made by a theory

Ideally, a theory should predict new evidence we’ve never even measured. Let me give you a famous example.

Einstein wrote his theory of general relativity in 1915. One of the consequences of his theory was that the sun would bend the light coming from other stars behind it.

But you can only measure this during a solar eclipse. Well, one came around in 1919. Astonomer Arthur Eddington took pictures of the eclipse and . . .

 . . .Einstein was right! Hot damn. Now that, my friends, is science.

But it gets better.

Another consequence of Einstein’s theory is the twin paradox. Take two identical twins. Leave one on earth. Send the other in a space ship travelling near the speed of light. When she returns to earth, the one that stayed on earth will be older.

It’s not that she’d look older or feel older. She would actually be older. She will have experienced more time.

Pfffff. Yeah right, Einstein. Nonsense. Gibberish. How can she have experienced more time? That’s the stupidest thing I’ve ever heard.

Guess what. Scientists tested this idea too. Einstein was right.

Holy Bat Balls in a Fryin’ Pan. Now that’s science.

Of course, they didn’t use identical twins, and they couldn’t travel at close to the speed of light. They synchronized two super-accurate atomic clocks, then put one on a plane and flew it around the world a few times, fast as they could. When they landed, the one that had stayed on the ground was ahead. It had experienced more time.

They’ve done this with all kinds of precautions and double-checking and different kinds of clocks. But Einstein was right. Sheesh.

Another example: astronomical cycles

A theory can even make a novel prediction about the past. In the 1930s, an astronomer predicted that cyclic changes in the Earth’s orbit could have caused the Ice Ages.

We didn’t have data about this at the time. Thirty years later, we came up with a way to measure ancient climates. Guess what? The history of the earth’s climate matched the predictions made by the theory of cyclic changes in the Earth’s orbit.

Evolution

Evolution, too, has predicted all kinds of new things not predicted by any other theory. We keep finding them to be true.

All these theories made specific predictions. They told us where to look for evidence, and left themselves open to being disproven. So far, these ones haven’t been disproven. Lots of others have been disproven.

That doesn’t mean these theories are proven. They could still be shown wrong at some time. But they fit the data better than other theories. And they’ve withstood thousands of attempts to disprove them. That’s pretty impressive.

Science performs controlled, repeated experiments

So far, we’ve seen 4 reasons why scientific knowledge is more reliable than other types of knowledge. It requires evidence, critical thinking, and falsification. And it prefers that a theory make novel predictions.

Now we look at another quality of good science that helps it provide reliable knowledge: controlled, repeated experiments.

When we “test” a theory’s prediction, that means we perform an experiment.

The theory and design of experiments could fill a library shelf. (So could each section of this article.) We’ll just take a brief look at how good experiments help us get at the truth.

How experiments work

Most experiments look for a relation between two variables. The one we think might be causing the other is called the independent variable. The one we think might be caused is the dependent variable. If we don’t know if either one is causing the other, we just call one the independent variable and the other one the dependent variable.

So, we take a small sample of subjects (people, plants, whatever) and run our test on them. Hopefully, the results of our sample reflect the whole population of people, plants, or whatever. For example:

  • We want to test which type of soil (independent variable) produces the most plant growth (dependent variable).
  • We want to test if religious belief (independent variable) affects the type of near-death experience (NDE) one might have (dependent variable).
  • We want to test if a smoking ban (independent variable) affects the rate of heart attacks (dependent variable) in a certain region.
  • We want to see if shoe size (independent or dependent variable) is related to reading ability (the other one).

Good experiments

Good experiments have a few qualities in common:

  • The independent variable is the only one that varies in the experiment. In other words, it is controlled.
  • The dependent variable truly reflects what we are studying. (It is valid.)
  • The variables can be measured accurately.
  • The experiment can be repeated.
  • The findings in our sample represent the whole population.

Controlled

Think about the plant growth experiment. Let’s say we put the same kind of seed in the same kind of pot but in different types of soil. We give all the plants the same water and sunlight and temperature. That’s what it means to have a controlled experiment. The type of soil is the only difference among our plants.

Let’s say we messed up and let them have different amounts of sunlight. Then some grow taller than others. How do we know if the extra growth was caused by the type of soil, or the amount of sunlight, or both? We don’t know.

So, we should control our experiments as tightly as possible. But there is a limit to what we can control. Should we make sure the guy with the watering can wears the exact same shirt every time he is watering the plants? Might that affect our results? We don’t know. It’s better to be safe than sorry, but we can’t control everything. We have to critique experiments on a case-by-case basis.

Another example. Let’s say we want to test if a certain drug makes people happier. So, we find some subjects. We measure their happiness. They take the pill. We measure their happiness again. If the second rating is, on average, higher than the first, our drug made them happier, right?

Not necessarily. Maybe just interacting with our friendly researchers made them happier. Maybe our warm lighting made them happier. Maybe putting anything in their mouths made them happier. Maybe they got happier as their anticipation grew for the $5 we're paying them to participate.

How do we control for these factors? We need half our subjects to be a control group. They will experience everything exactly the same as the other people, except for the drug. They’ll take a pill with our friendly researchers, under our warm lighting. They’ll swallow the pill. They’ll get paid $5.

But the control group will take a pill with only sugar in it, not the happy drug. But they won’t know that, and neither will the people giving them the drug. Only the old guy in the back room doing the math will know who got what.

When we look at the data, let’s say the people who got the happy drug were much happier afterward. People who took the sugar pill were only a little bit happier afterward. Assuming we took a random sample of people, nothing else was different—on average—between the two groups. Now we know the drug actually worked.

Or at least, we think so. We should check our data and run the experiment a few times to be sure.

This is the kind of thing required by the FDA before they let a new drug on the market. This takes many years and millions of dollars, and that’s why drugs are so expensive. That and the risk a company takes in getting sued over their drug—no matter how much testing they do.

If you want cheaper drugs, either (a) stop suing, or (b) be wiling to take less well-tested drugs.

Validity

An experiment can also be a measurement, but that’s a little different. We have less control. But we can still do our best.

Let’s say we want to know the age of a manuscript. Because the old manuscript was made from a tree (a plant with carbon in it from the atmosphere), we could use carbon dating. That’s what they did with the Dead Sea Scrolls.

With carbon dating, we are actually measuring the amount of C-14 (a carbon isotope) in the pages. Usually, this can tell us about how old they are. To see how, read this.

But we have to be careful. When we measure C-14, are we really measuring the document’s age? Is our measurement valid?

There are ways to check this. One is to take many samples, many measurements. If they all agree, that makes us more confident.

Another way is to check for signs of contamination. There are even ways of “cleaning away” C-14 contamination in the lab. We can still try to get a good measurement.

Another way is to compare our results to the results of other dating methods. We could try textual analysis. Does the manuscript mention an earthquake we know happened in 300 A.D.? If so, it wasn’t written before 300 A.D.

Scientists do their best to make their experiments and measurements valid. They want to make sure they’re measuring what they think they’re measuring.

Measured accurately

Obviously, we want to make sure we measured the plant growth or happiness levels or rate of heart attacks accurately.

About that heart attacks experiment. Officials enacted a smoking ban in Helena, Montana from June–December in 2002. The town is isolated and has only one hospital. During the ban, heart attacks at that hospital dropped 60%.

It’s pretty easy to check that. Just look at the records. But you need to be sure the hospital was recording heart attacks accurately.

Repeated

It’s best if an experiment is repeated. Many times. This eliminates error. Maybe there was a mistake in one experiment. Maybe it didn’t use a representative sample. Maybe it was a fluke.

But if we get the same results many times, we can be more sure.

Measurements work the same way. If we repeat a measurement many times and keep getting the same answer, that increases our confidence. We’ve measured the position of Jupiter lots of times. It always turns up where the theory of gravity says it should be.

Representative

Back to the happy pill. The ones who took the drug were much happier than the ones who took the sugar pill. Does that mean we should sell the pill across the country?

We need to know if our sample can really represent, or “stand in for” all people.

That’s why we took a random sample. If we just recruited people through the Internet, that would be a biased sample. We’d only be testing regular Internet users. Maybe people who tend to use the Internet have some trait that causes the happy pill to work better on them than other people.

Or maybe we just recruited in northern Minnesota, so our entire sample was white. Maybe other races have some trait that would cause the happy pill not to work on them. Just because it worked on our sample of white people doesn't mean it will work with most people.

Or maybe there’s some factor we aren't even thinking of! That’s why we need a random sample, so there is an equal distribution of every factor—even ones we don’t know about.

We need a big sample, too. Test only 10 people and a fluke can skew our results completely.

Example of a bad experiment

Bad experiments don’t have these traits, so their results are not reliable. Let’s look at one example.

In 1982, Dr. Randolph Byrd set out to prove that prayer to the Christian God can improve hospital recovery. 393 incoming patients agreed to take part in a randomized study. A computer assigned half the patients to be prayed for by “three to seven” Christians. The other half were not assigned any intercessors.

At admission, there were no average differences between the two groups. 29 factors were measured during their hospital stay, for example incidence of pneumonia and length of hospital stay.

At the end of the study, Byrd found some differences between the groups and concluded that “there seemed to be an effect [from prayer], and that effect was presumed to be beneficial.”

Let's look closely for problems with the experiment. This is what happens to every study. It’s called “peer review.”

Byrd admits the first problem right away: “It was assumed that some of the patients in both groups would be prayed for by people not associated with the study; this was not controlled for . . .”

Read that again. The very thing being studied—prayer—was not controlled for! That alone can invalidate the results. That’s a big mistake.

You might say, “Well, how was he supposed to keep people outside the study from praying for the patients?” I don’t know. That would be really hard. Maybe impossible. But that doesn't make the experiment any better. That just means it’s a really hard topic to study scientifically. There are lots of topics like that.

Second, there are some theoretical problems. If Byrd’s results were reliable, what would that say about God? Is he a cosmic vending machine, forced to offer help to whichever side gives the most prayers? Does he not have compassion on those who get fewer prayers? The results wouldn't even fit the theory (the efficacy of prayer to the Christian God) that Byrd is trying to prove!

Third, let’s take a closer look at Byrd’s data. Among the group that received dedicated prayer:

  • 5 percent fewer needed diuretics.
  • 7 percent fewer needed antibiotics.
  • 6 percent fewer needed respiratory intubation and/or ventilation.
  • 6 percent fewer developed congestive heart failure.
  • 5 percent fewer developed pneumonia.
  • 5 percent fewer suffered cardiopulmonary arrest.

No other differences were found in the other 23 variables. Including death. And despite specific prayers to prevent death.

As for the other 6 variables, even Byrd admits the differences “could not be considered statistically significant.” Then why did Byrd even report them? Why didn’t he just say, “No significant differences between the groups were found”?

Byrd says he overcame the statistical insignificance with a severity score and a multivariate analysis. But neither method is meant to deal with interdependent variables. Look at those 6 variables. They influence each other. They are interdependent.

Develop pneumonia, and you need antibiotics. Develop congestive heart failure, and you need diuretics. The variables are dependent on each other. Byrd’s methods can't deal with that. Those 6 variables, along with all the others, really showed no difference between the two groups.

Gary Posner illustrated the problems with this study here. Here’s my own illustration:

Say we want to test the effect of reading porn on hospital patient recovery. One group of patients are given 3–7 “dirty mags” each day by the scientists. The other patients are given none. Patients in both groups are visited by friends and family, who bring some of them more porn (but we don't know who or how much). Across 29 variables, no differences between the groups are found, unless we use inapporpriate methods on a few interdependent variables.

Would we conclude that “There seemed to be a beneficial effect of porn?” Of course not.

(Good) scientists take great care to protect the truth of their findings. That’s why they conduct very careful experiments and measurements, and submit them to peer review. And that’s why we trust (good) science much more than less controlled methods.

Science prefers simple explanations

Why do scientists assume that the universe is simple? They don’t. It’s just that simple explanations tend to be more likely than complicated ones. There are fewer potential problems with a simple theory.

We use the same logic in court cases. Let’s say Jacob is on trial for murder. He has no alibi for the time of the murder. His fingerprints are on the gun used to kill a girl. The prints from his boots were in the mud next to the dead girl.

The jury needs to decide what explanation of the facts is most likely. Here are some possibilities:

  • Jacob killed the girl.
  • Jacob’s neighbor Lyle killed the girl. He used special tape to transport Jacob’s prints from his home to the gun. He stole Jacob’s boots and wore them that night, then cleaned them and snuck them back into Jacob’s closet.
  • An alien from outer space killed the girl. Its skin can mimic any fingerprint or footprint. The alien disliked Jacob, and framed him for the murder.

Those are interesting possibilities, but I think we can agree the jury should convict Jacob of murder.

Unless there is other evidence. Maybe there is independent evidence that Lyle took Jacob’s prints and boots. Maybe Lyle left one of his socks in a boot. Maybe a security camera recorded Lyle wearing John's boots near the crime scene that night.

But if we don’t have independent evidence like that, the simplest explanation is the best. The others are possible, but they are not the best. There are too many details with the others that we have no evidence for.

Science is tentative and progressive

Scientists know the limits of human knowledge. They don’t (or shouldn’t) claim to have “proved” something. They don’t have “absolute” knowledge. Science is tentative.

Science is also progressive. Our current theories are not complete or perfect, but they’re closer to the truth (they work better) than the theories we had 50 years ago. Einstein’s physics is better than Newton’s physics, which was better than Aristotle’s physics, which was better than ancient Hindu physics.

Some people say, “If science was reliable, it wouldn’t keep changing all the time.” Actually, science is reliable precisely because it keeps changing—in one direction: towards truth.

Science corrects itself. When a new theory works better than an old one, we abandon the old one.

When Big Bang theory was proposed, most astronomers didn’t accept it. They simply couldn’t believe the universe had a beginning. “Big Bang” was actually a sarcastic term for the theory from one of its critics, Fred Hoyle. But the evidence kept mounting up in favor of “Big Bang” theory, so scientists slowly began to accept it. Because of the evidence, it’s now the dominant theory.

But it’s still up for debate. There are lots of other theories. If one of them gets more evidence than Big Bang theory, it will become the dominant theory.

Summary

Trust me, that was a whirlwind tour of these issues. You could spend a century reading about them and debating them.

Here’s what we saw:

  • Science is a much more reliable path to knowledge than any other method . . .
  •  . . .because of the methods scientists choose to use. They are strict, reliable, and careful methods.
  • That is why we trust science. Not because it’s perfect. But because it’s way better than other methods.

Of course, we can’t just say that “geology is a science, so everything geology finds is reliable knowledge.” No, we need to evaluate geology’s findings on a case-by-case basis. Lots of things claiming to be science aren’t good science.