Unlocking Magic (part 2: under science’s watchful eye)
Mixing magic with technology can be a challenge. We’ll explore some of science’s fundamental tools to see if we can unlock some magic underneath it’s gaze.
This post comes slightly out of order from my plan for the Unlocking Magic series. I’d planned on a continued slow descent from the 10,000 foot view of magic down into the weeds of how to actually make a magic system work. However, a recent question on Worldbuilding gives me a chance to dive right into what Brandon Sanderson would deem “soft magic,” and show it off a little. Soft magic is uniquely well suited for appearing in technologically advanced worlds because its lack of hard rules avoids conflict with the highly rule oriented nature of science and technology. It is designed to bring awe and wonder to the world subtly, and subtlety can be a powerful tool in the hands of a world builder. Without further ado, I would like to present a few ways to blend soft magic and technology in a way that ensures one does not step on the others’ toes.
When scientists put together their experiments to test a theory, there are a few fundamental requirements. If these requirements are not satisfied, science has a difficult time using the scientific method to gauge the validity of its theories. The two most important requirements are that the experiments are repeatable and reproducible. A repeatable experiment is one which, if performed again in the same environment, potentially by the same scientist, yields consistent results. A reproducible experiment is one which yields consistent results if someone else does the test in a different environment. These requirements can be endlessly useful for weaving softer magics in with science: any soft magic which is not repeatable or reproducible is simply not interesting to science! You can fit an endless quantity of soft magic alongside science this way.
We can explore the meaning of these two powerful words with some examples. We are all familiar with tests involving rolling a ball down an inclined plane to measure the acceleration of gravity. This test is repeatable. If you were to do the experiment 10 times, and then do it 10 more times, the results are generally statistically similar between the first group and the second. The test is also reproducible: every student who does this experiment in physics class arrives at results consistent with the results of the other students. This experiment to measure gravity is repeatable and reproducible, and thank goodness for that! (Grading lab reports as a TA would be brutal if it wasn’t!)
Scientific tests often fail at either repeatability, reproducability, or both. Non-repeatability happens all the time due to statistics. Sometimes you just got a statistically unusual set of results and you never get it again. Other times it arises because you, yourself, influenced the results. Psychology studies take great care to make sure the 100th test subject gets exactly the same treatment as the 1st test subject. They will do things like blinding themselves to the data until after the experiment is over to prevent partial results from skewing the way they treat the next subject as they introduce the test to them.
Non-reproducability is very frustrating, especially for so called “crack pot” scientists whose results never seem to be reproducible by anyone besides themselves. It can also show up when a scientist accidentally biases the experiment. A scientist exploring the tossing of dice might find unusual results if they unconsciously “helicopter” the dice. Helicoptering the dice is a technique used to influence the dice at a Craps table (and is, understandably, illegal at most casinos). These results will be repeatable, but not reproducable. Another scientist, who does not unconsciously helicopter the dice, will observe a very different result. (In science, the only difference between the “crack pot” scientist and the scientist who unconsciously influences the experiment is that the latter will work with the community to refine the experiment to reduce this bias, while the “crack pot” will try to convince everyone that his results are the only right ones.)
So we have a region for our soft magic which science cannot easily disprove: it must be either non-repeatable or non-reproducible if it wants to avoid science’s critical eye. It may even use a little bit of both, leaving questions in its wake, being slightly repeatable and slightly reproducible. There are myriad ways to blend these into your magic, so please use your own creativity to figure out the right path for your own world. That being said, I’ll go through a few real world examples to help get the juices flowing.
The easiest way to dodge science’s sharp gaze is to have a “fickle” magic. If your magic never does the same thing twice, it becomes much harder to study. How hard? In science, we often announce results with respect to a P-value, which is a measure of confidence. The smaller the number, the less likely we think the results are just statistically unusual (alternatively: the more likely we think the results are meaningful). For particle physics, where we can make particles do the same thing billions and billions of times, it is not uncommon to see P-values of 0.000001, indicating an uncertainty of one part on a million. Now consider a realm where the test subject is more fickle: psychology. In psychology, a P-value of 0.05 is often acceptable, because science simply can’t control people enough to get a more statistically relevant answer! Science becomes more and more uncertain of its results as the subject demonstrates some level of individuality or will.
If you have a “living” source of magic, whether it is a life-energy or a fae magic or any other magic with its own personality, science’s control of it is limited to its ability to manage this “fickleness” in a statistically significant manner. Your mages, however, don’t need it to work statistically… they need it to just work for them. If their source of energy lets them know a fireball is impossible, but an icestorm is fair game, the mage can flex. Science has a much harder time flexing in that respect without creating repeatability or reproducability issues.
We see this exemplified in ghost hunting. Modern ghost hunters have all sorts of fancy scientific gizmos, but they are studying a living source of energy. The ghosts do not always play nice (or play at all). Accordingly, the quality of their scientific results is often lacking. There’s too much they can’t control, and in many cases are too polite to control.
There is another, darker, way to control the statistical powers of science. What if you cannot reproduce an experiment because you don’t want to? This is a route dark magics often take. If you look at many dark magics, they rely on you to pledge something very dear and un-reproducable, often your own life. With a sample size often as small as 1, science cannot draw any useful statistical results. With dark magic like this, we often see the most powerful of mages seeking to convince test subjects to unwittingly bid their own life so that the mage may observe the outcome of their demise. The mage seeks scientific answers but refuses to give them to others. And, of course, any experiment that leads other mages to come and kill you for your experimentation puts a damper on science, just as Galileo found as he argued against the Church (though he was merely put under house arrest for the remained of his natural life serving a commuted sentence).
(Enough of that dark brooding stuff. Hey look! Rainbows! Butterflies! Arbitrary distractions! — Oh, hey! Look! Individuality!)
One of the easiest ways to weave science and magic into the same world is to make magic personal. If magic works a little bit different for each person, no test of said magic can be reproducible. Consider, for a moment, the age old art of visualizing a victory. Science has done their studies and found that people are statistically more likely to succeed at a great feat if they visualize it beforehand, because that is a repeatable and reproducible experiment. But what should you visualize? That is a much harder question for science to answer, because it is so personal. The best visualization differs from person to person. When Michael Phelps began training, he was told to visualize the perfect race every day. Nobody told him the details of what to visualize. He was just told to find his perfect race, and visualize that. If you were to try to scientifically study the particular details his vision included (like the feel of the wake behind him), you’d find the effects were not reproducible. They’re magical, bound only to him.
This approach is visible in real life in many martial arts, most famously associated with Kung Fu disciplines of China. While each school has its curriculum which every student is expected to master, the most advanced pupils are guided towards developing their own particular style within the structure of the school. Such practitioners may custom tailor their style to their unique body to do things that appear otherwise impossible (such as one with a particularly flexible hamstring striking with impossibly high kicks). Is it magic? Is it physics? Science is not in a good position to tell the difference because this personalized style is not reproducible by anyone else. It can axiomatically declare “magic doesn’t exist, so what he’s doing must be physics,” but it can’t actually prove anything about it unless the Kung Fu master in question permits science to take his magic apart.
(factoid of the day: Kung Fu is not actually a reference to martial arts. It is applied to any skill which takes time, energy, and discipline to acquire. When Jackie Chan says “Kung Fu lives in everything we do, he isn’t just talking about martial arts.”)
This personal approach proved to be the most useful for the World Building question that inspired this blog post. Their particular concern was how to make sure magic was interesting enough to be worth including in the world, but not powerful enough to replace technology. They wanted the “big” world to be dominated by technology and the “small” support a thriving magical nature. Many of us have been on a team; some have even lead a team. Those who have practice leading a team in the modern workplace knows just how hard it is to get a large number of strong minded individuals to work towards a common goal. A personal magic permits us to use this real-life effect to solve a world building problem. While technology permits us to add small pieces over and over to grow arbitrarily large, adding wizards to a team is not so simple. This limits the power of wizards to small things that one or two wizards have power over, while technology builds cities and bridges and railroads. This system is also beautifully balanced. Should technology ever act so egregiously as to offend all wizards at once, suddenly technology would find itself on the receiving end of an unexpectedly unified wizard front!
This leads me to the final topic for today, an advanced world building approach for the mathematically inclined. You can build an entire world without dropping to this level of mathematical rigor, but it happened to be very useful in the world described above. If you spend enough time as an engineer, you learn to love “normal” distributions. The “normal” or “Gaussian” distribution is a distribution of a random variable that shows up all over science and technology. In fact, most scientific hypotheses are tested against the assumption that all errors are “normal.” Why is this such a popular distribution? Well, for starters they’re easy. Normal distributions interact with each other in a really sane way, permitting us to predict how a structure will perform long before we build it.
Of course, all of this study is pointless unless we can devise situations where the real world is well modeled with normal distributions. That should be hard, right? Well there’s a neat theorem called the Central Limit Theorem (CLT). CLT basically states that the more random samples you average together, the more the expected result looks like a normal distribution. Its an incredibly fundamental tool for scientific analysis. It’s used to explain why when you pick up a glass of water to drink, you don’t have to consider the exact movement of all of the atoms in the glass. There’s enough atoms that the average motion is pretty much… well… the motion of a glass getting picked up.
There are a few limits you can attack for sneaking magic in underneath technology. Along the lines of the repeatably arguments before, you can prevent CLT from working if the samples are not independent. CLT is only valid for independent draws (this causes all sorts of shenanigans in simulations when your model assumes two variables are independent when they are, in fact not). In fact, the “fickle” magic from before may draw upon this. If one spell affects the next, science cannot use the CLT because the spells are no longer independent.
There is another funny limitation to the CLT which is applicable to the world with science and technology from above. The CLT only works with distributions that have a well defined expectation and a well defined variance. Those are statistical terms, so you won’t be able to take advantage of this limitation without some background in statistics. However, if you’re willing to deal with infinite expectations or infinite variances in your magic system, you can actually develop a truly hard magic system where all of the rules are defined mathematically, and still have it operate side by side with technology, simply because such magic prevents the CLT from being applied, and most of science and technology relies on it.
What does that mean without the math speak? If you find yourself in a situation where you might say, “If only they could just put aside their difference and unite, there’s nothing they couldn’t do,” you are in an excellent position to uproot a science and technology based villain. Wordings like that suggest, in that statistical speak above, an infinite variance. This means the villain, applying the CLT like any good science and technology practitioner will reliably underestimate the heros if they unify!
Of course, there’s no real reason why magic and technology need to be separated in the first place, so feel free to let them mingle. In future articles, I plan to delve into ways to do that without the result being hard to relate to. In the mean time, I would be remiss if I did not close with Arthur C. Clarke’s famous law, to leave you to ponder and explore:
“Any sufficiently advanced technology is indistinguishable from magic” — Arthur C. Clarke
Happy world crafting!
-C
