 A friend who's in liquor production has a still of astounding construction. The alcohol boils through magnetic coils. He says that it's proof by induction. Human beings are, for better or worse, stuck in a subjective experience of the world that can vary in important ways from person to person. We all generally agree that there's something out there that we can all touch and experience in our own unique ways. But how much of that experience is us, and how much of it is actually out there? Accessing truth is something that we all do, or think that we do, all the time, using all sorts of different methods that we can choose to trust more or less than the others. There's philosophical rationalism, where the truth is revealed if you reason hard enough. There's religious faith, where the truth is revealed through revelation, dogma, or holy texts. There's intuition, or trusting your gut. If you're watching thunk, you probably trust the claims of science and scientists a great deal, even knowing that the romanticized ideal of SCIENCE conflicts with the mundane reality of spending years staring at columns of data in Excel while worrying about funding. But in some debates, where somebody is disagreeing with scientific consensus, whether they're anti-vaxxers, or creationists, or climate change deniers, I have heard some proponents of science assert that it's 100% objective, that unlike other methods of accessing truth, it's a perfect abstraction of logic and evidence that's totally free of any bias or subjectivity. Well, not really. Most implementations of the modern scientific method make a number of important assumptions that are, on some level, arbitrary. Whether or not they're as arbitrary as assuming that your gut is never wrong is up for debate. But they are still assumptions, and they are still arbitrary, and knowing what they are is an important part of understanding science and its operation. For example, in order to do science, one has to make an implicit assumption of naturalism. That is to say that all phenomena are caused by natural causes, not supernatural ones. This is why papers entitled thermal fluctuation in transport in type 2 superconductors, or maybe fairies are playing with our gauges, don't get published. That assumption isn't logically bulletproof, but it is necessitated by science's requirement of experimental repeatability. If the data that you're collecting today is subject to the whims of some supernatural entity, which you can't measure or control for, who's to say they won't be totally different tomorrow? That's not to say that someone can't be a good scientist and still believe in the supernatural outside of the lab. It's just that when they're doing science, they generally must make the assumption of purely natural causes. There's also a general assumption of the universal homogeneity of physical laws, that the rules which govern the things in your experiment are just as applicable outside the lab, that they're relatively unchanged by time or space. I mean, science would be pretty unremarkable if all it did was tell you things that happened in labs, and that's not just because a full day of pipetting is really boring. There's another assumption that science makes which isn't really intuitively obvious just because we take it for granted so often in our lives, induction. Induction is the process of recognizing patterns and then generalizing those patterns into rules and is generally considered to be a foundational component of scientific research. It's also a problem, so much so that there is a capitalized problem of induction. Let's say that you have two billiard balls on a pool table. You roll the first one into the second one a hundred times and each time the first one stops and the second one starts moving. You'd probably say that the first ball's motion starts the second ball rolling and that other billiard balls will behave in the same fashion. You might even write an equation to describe how the second ball will move. The capital P problem, most notably raised by David Hume, is that no matter how many times we see that same sequence of events, there's no concrete reason to expect it to repeat. You've probably heard that correlation doesn't imply causation, but unfortunately nothing necessitates it. We can't see causation. It isn't some substance that we can interact with. We can only see one event followed by another event and hope that they keep on happening in that sequence in the future. Maybe on the hundred and first trial, the first ball reverses direction or vanishes all together. There's no guarantee that those things won't happen, we're just guessing that they won't. And unfortunately we can't say that induction just happens to be right a lot of the time and justify it that way because that's using induction to justify induction. That's a problem, at least for philosophers and especially for philosophers of science. Without a concrete justification for induction, all of the capital T truths that we think we know due to science are really just hypotheses waiting to get knocked over by one contrary observation. And that's kind of the reality of the situation. When CERN announced the existence of the Higgs boson, some people celebrated the discovery as though somebody at CERN had grabbed one with a pair of really good tweezers and held it under a microscope. But in reality, in order to establish the existence of the Higgs, CERN had to use some very sensitive equipment and in the paper announcing the discovery, there was a value, the same value that's in many modern scientific papers, which stated the odds that the data that they hoped represented the Higgs was just a fluke. I mean this is one of the most significant discoveries in physics to date and it's got this asterisk attached to it that says there's still a 1 in 3.5 million chance that if there were no Higgs, we'd still have gotten these results. It's not a certainty, it's still just a guess. So science makes a number of arbitrary assumptions not unlike other methods for accessing truth. And because of the problem of induction, we can't even be sure that the things that we supposedly know due to science are actually true, they're just more probable than not. Why is it so popular? Well, because it's remarkably successful, despite the fact that science isn't some perfect objective entity abstract from human thought, despite the fact that scientists aren't any less prone to bias than anybody else, despite the fact that scientific discoveries are just statistical likelihoods, not necessarily true, science is still way better than any contender at making very accurate, very specific, and very risky predictions. Out of all the methods which people use to access truth, science isn't different because it's perfectly objective or free from assumptions, there's just as many assumptions in science as anything else. But those arbitrary assumptions still seem to be the best ones that we have for saying there will be an eclipse at this time, or if you smash these particles together at this speed, then you'll see an energy spike at this frequency and that's the Higgs boson, probably. Whether that's because they're more accurate than other assumptions about the universe, or because they're just better suited for making accurate predictions, that's a different question. Whether someone wants to accept those assumptions as actually true, or just necessary for getting stuff like medicine and computers, that's their call. But I've at least got to respect the fact that out of all the methods of accessing truth available to me, only scientists can even venture a guess at how many inches of snow I'm going to have to shovel off of my driveway. Do you think that the assumptions inherent to science are true, or just necessary? Do you live on Long Island and want to lend me a snowblower? Please leave a comment and let me know what you think. Thank you very much for watching. Don't forget to blah blah subscribe, blah share, and don't stop thunking.