What Flat Earthers Can Teach Us About How Science Works

If you didn’t know, there is a growing community of “flat earthers” who believe:

1) that we all live on a flat land mass (not a spinning globe traveling around the sun), and

2) that governments and scientific organizations have conspired to hide this fact.

They have their own conferences and podcasts. Some members of the community build miniature flat-earth models to demonstrate how a flat earth would work. And there are — just as in professional scientific communities — bitter controversies within the community, both personal and theoretical. I learned most of this from the documentary Behind the Curve, available on a bunch of streaming services.

People on the internet love to make fun of flat earthers. And clips from the film a few years ago helped people do just that. The main experiment performed by flat earthers featured in the film became a widely shared clip because it seems to either “prove” the earth is round or “disprove” the earth is flat. And, at least as portrayed in the film, the flat earthers continue blithely on in their belief of a flat earth, rejecting their own evidence.

But the idea that the flat earthers made some sort of fundamental reasoning error when reflecting on the results of the experiment is a mistake. It reflects a serious misunderstanding of how science works.

The work put in by the community to design and reason about the experiment was an example of good scientific reasoning, not bad reasoning. Their ultimate conclusion that their evidence wasn’t strong enough to change their minds is typical of actual scientific practice and not a bad thing at all. As the film points out — believing in a flat earth is not just (or even mainly) about science — but to the extent that it is, the community’s reasoning errors are more subtle than “LOL, they just proved themselves wrong!”

Let’s explore what the experiment was, and why it wasn’t a mistake to reject the evidence they gathered.

The film documents members of the flat earth community performing experiments meant to establish that the earth is not round. The main experiment involved stretching a beam of light over a long distance. If the earth is flat, they reasoned, then the beam should remain at a consistent height above the ground. If the earth is round, then the beam should be closer to the ground at a point distant to it. My crude depiction is below.

This is a great example of scientific reasoning. They had to:

1) Consider two potential models of the earth. In one, the earth curves. In the other, the earth is stationary and flat; the stars, sun, and moon rotate on top of a pancake-like earth.

2) Have some sense of the consequences of each model. When we compare models to each other, often both models will make some similar predictions, but there will also be areas where the models differ.

3) Come up with a reasonable test that would distinguish these two models. Note how the test they choose did not tackle all aspects of either model. It tackled just one specific difference: the effect earth’s curvature on a straight line.

Their initial idea was to use a powerful laser as the light source. But they ran into trouble: the light from the laser disperses at far distances. The main experimenter had considered this issue and made modifications to the laser to narrow its beam. After unsuccessful field tests, he modified the approach. In the next iteration he used a spotlight and other team members positioned large boards with holes cut out in them at different intervals. After considerable work, they pull the test off. As he points out in the film, running the experiment gave the principal investigator a sense of just how hard experiments really are.

So far, so good, right? All of this is normal scientific practice. You have some idea that you’re interested in testing. The idea implies something about the world that you can actually measure. You figure out how to measure it precisely. You spend some time testing your measuring instruments, modifying them as necessary. And, with any luck, you have an observation or test that shows you the information you were looking for.

So what happened? When raised high, the light can be seen on the far side, but when lowered it couldn’t be seen. Huh. That’s a problem for the flat-earth view because it suggests that the curvature of the earth might be blocking the light. But the flat earthers had an explanation: weeds were getting in the way, which prevented the light from making it’s way through when the spotlight was lowered. In other words, it wasn’t the curve of the earth that was preventing the light from getting through, it was the damn foliage.

“Ah…,” you say, “See!” “The experimental set-up was fine, but the problem was in what they did after the experiment: their experiment proved the flat earth theory wrong but they couldn’t accept this!” But this is also incorrect. There’s nothing about the flat earthers’ explanation for their experimental results that is inherently unscientific.

Conceptual change in science does not come from a single piece of evidence; it comes from the accumulation of evidence over time. The Michaelson-Morley experiments were the first to suggest that, unlike everything else, measurements of the speed of light do not vary with the observer’s speed or perspective. So scientists did one experiment and called it a day, right? No, of course not. Lots more experiments confirmed this mind-boggling idea. And Einstein developed an incredibly powerful model by accepting the accumulation of experimental evidence that further explained several other perplexing observations (like Mercury’s unusual orbit).

Yes, looking from the outside, it can seem silly to “blame it on the weeds”. But scientists are constantly faced with this problem. Is the instrument calibrated correctly? Is there something interfering with the reading? Is there some other explanation that I haven’t thought of?

Single experiments never definitively prove (or falsify) any hypothesis. That’s because experiments never test single hypotheses. Surprising results in science come up often. Results that, you might think, “falsify” or “disprove” the target hypothesis. But it’s often quite reasonable to maintain the target hypothesis and ignore, reject, or reinterpret the data instead. Professional scientists do this all the time.

The critical part is to keep going. It’s not enough to perform one experiment. Is there another way we can test the same idea? Could we repeat the experiment with the same protocols? Do the protocols need updating? What explanation accounts for all of the evidence — experimental results as well as other things we know? A reasonable next step would be to try to make the experiment more precise, so that the weeds explanation would have less force (maybe it was weeds one time, but two times, three times, four times? Maybe it was the weeds in this area, but we’re on a flat plain or desert now.).

Public discourse on science tends to focus on single experiments. The media often reports results from single studies and acts like the results are revelatory. This approach is rooted in a hypothetico-deductive model of science (“science as logic”) that is almost as far from an accurate description of science is as a flat earth model is to the actual earth. Science is about a lot of things — measurement, explanation, argument, replication, modeling, and more — but it’s certainly not about how single experiments prove something true.

Believing in a flat earth is not primarily the consequence of bad experimenting. We think, “their belief is incorrect (and even, to us, absurd), so they must be evaluating evidence incorrectly.” But, at least in their experimental work, there’s nothing unusual about the way they go about it. It’s just that they’re starting from a very different position than we are. They didn’t get to “flat earth” through careful experiments and scientific reasoning (most of us weren’t introduced to the idea of “round earth” that way either — we were just told that that’s the way it was). From what I can tell, flat earthers came to “flat earth” because they favor conspiratorial explanations of the world and wanted to join a social (and supportive) community. One of the main figures profiled in the film talks about how he was looking for a conspiracy to believe in.

So are flat earthers making a scientific reasoning error? Yes, but it’s not about performing experiments — it’s about evaluating explanations. Thoughtful people have accepted that the earth is round for several millennia. Why? Because a round earth neatly explains everyday observations and a flat earth cannot.

Consider the following:

• When boats approach a mountain, the mountain appears to rise from the ground (instead of slowly appearing all at once). This was noted in the West by Ptolemy, the influential astronomer from the 100s A.D.
• Similarly, the bottom of boats far on the horizon cannot be seen, due to the curvature of the earth.
• Similarly, when boats sail away from shore on a clear day, their hulls disappear first, before their sails.
• As you travel south, southern constellations appear to rise from the horizon. This was noted by Aristotle in the 300s B.C.

In modern life, we may not have had these experiences. But we can certainly go experience them if we want. We moderns, however, have another set of confounding questions and observations that conflict with the idea of a flat earth:

• A flat earth presumes there’s a perimeter around the earth. Pick any perimeter you want. How can you fly between two points, which are supposedly on opposite ends of the earth, so quickly?
• Most flight paths will look curved when presented on a flat screen or map. Why? A round earth provides an explanation — curved paths are often the shortest routes — while a flat earth does not.
• How does GPS technology work on a flat earth?
• How do satellites work on a flat earth? Is it impossible to get to space? Are they just floating up there?
• For that matter, how does gravity work on a flat earth?

It’s not about “proof” — it’s about explanatory coherence. When a single explanation makes sense of existing evidence, especially lots of different kinds of evidence, then the scientific thing to do is to hold onto that explanation. If that explanation can predict and control future events — like how we can build an entire GPS system based on modern physics and it works — that’s even stronger evidence that the explanation is, at the very least, extremely useful, if not straight-up true.

Of course, it’s another form of explanatory coherence that keeps conspiratorial thinking going: lack of evidence for the conspiracy is only further evidence of a conspiracy because the conspirators clearly hid the evidence.

Experiments are probably not going to change many flat earthers’ minds. There’s too many other reasons for them to believe: believing means belonging to a supportive social group; not believing means being, in some sense, ostracized from that group. Believing means feeling like you know something most people don’t; not believing means accepting that you’ve been (quite publicly and loudly) wrong. But if we believe in the power of scientific research to change people’s minds — and, I do — then we should be encouraging their research. Not ridiculing it.