“You cannot observe an imaginary number in reality.” Have you ever heard this claim before? It has a nice ring to it, and sounds almost tautological. I’ve personally heard the claim made by several professors of physics, alongside a host of less qualified people. So let’s take a close look at it and see if it holds up to scrutiny. *Can* you in fact observe imaginary numbers in reality?

First of all, let’s ask a much simpler sounding question. Can you ever observe a real number in reality? Well, yeah. Position, time, charge, mass, and so on; pretty much any physical quantity you name can be represented by real numbers. But if we’re going to be pedantic, when we measure the position of a particle, we are not technically *observing a number*. We are observing a position, a physical quantity, which we find has a useful *representation* as a real number. Color is another physical phenomena that is usually represented mathematically as a real number (the frequency of emitted light). But we do not necessarily want to say that color *is* a number. No, we say that color is a physical phenomena, which we find is usefully described as a real number.

More specifically, we have some physical phenomena whose *structure* contains many similarities to the *abstract structure* of these mathematical objects known as real numbers, so it behooves us to translate statements about the phenomena over to the platonic realm of math, where the work we do is precise and logically certain. Once we have done the mathematical manipulations we desire to get useful results, we translate our statements about numbers back into statements about the physical phenomena. There are really just two possible failure points in this process. First, it might be that the mathematical framework actually *doesn’t* have the same structure as the physical phenomena we have chosen it to describe. And second, it might be that the mathematical manipulations we do once we have translated our physical statements into mathematical statements contain some error (like maybe we accidentally divided by zero or something).

So on one (overly literal) interpretation, when somebody asks whether a certain abstract mathematical object exists in reality, the answer is always no. Numbers and functions and groups and rings don’t exist in reality, because they are by definition *abstract* objects, not concrete ones. But this way of thinking about the relationship between math and physics does give us a more useful way to answer the question. Do real numbers exist in reality? Well, yes, they exist insofar as their structure is mirrored in the structure of some real world phenomena! If we’re being careful with our language, we might want to say that real numbers are *instantiated* in reality instead of saying that they exist.

So, are imaginary numbers instantiated in reality? Well, yes, of course they are! The wave function in quantum mechanics is an *explicitly complex entity — *try doing quantum mechanics with only real-valued wave functions, and you will fail, dramatically. The existence of imaginary values of the wave function is *absolutely necessary* in order to get an adequate description of our reality. So if you believe quantum mechanics, then you believe that imaginary numbers are actually embedded in the structure of the *most fundamental object there is:* the wave function of the universe*.*

A simpler example: any wave-like phenomena is best described in the language of complex numbers. A ripple in the water is described as a complex function of position and time, where the *complex phase* of the function represents the *propagation* of the wave through space and time. Any time you see a wave-like phenomena (by which I mean any process that is periodic, including something as prosaic as a ticking clock), you are looking at a physical process that really nicely mirrors the structure of complex numbers.

Now we’ll finally get to the main point of this post, which is to show off a particularly elegant and powerful instance of complex numbers applying to physics. This example is in the realm of electromagnetism, specifically the approximation of electromagnetism that we use when we talk about circuits, resistors, capacitors, and so on.

Suppose somebody comes up to you in the street, hands you the following circuit diagram, and asks you to solve it:

If you’ve never studied circuits before, you might stare at it blankly for a moment, then throw your hands up in puzzlement and give them back their sheet of paper.

If you’ve learned a *little* bit about circuits, you might stare at it blankly for a few moments, then write down some complicated differential equations, stare at them for a bit, and *then* throw your hands up in frustration and hand the sheet back to them.

And if you know a *lot* about circuits, then you’ll probably smile, write down a few short lines of equations involving imaginary numbers, and hand back the paper with the circuit solved.

Basically, the way that students are initially taught to solve circuits is to translate them into differential equations. These differential equations quickly become *immensely* difficult to solve (as differential equations tend to do). And so, while a few simple circuits are nicely solvable with this method, any *interesting* circuits that do nontrivial computations are at best a massive headache to solve with this method, and at worst actually infeasible.

This is the real numbers approach to circuits. But it’s not the end of the story. There’s another way! A better and more beautiful way! And it uses complex numbers.

Here’s the idea: circuits are really easy to solve if all of your circuit components are just resistors. For a resistor, the voltage across it is just linearly proportional to the current through it. No derivatives or integrals required, we just use basic algebra to solve one from the other. Furthermore, we have some nice little rules for simplifying complicated-looking resistor circuits by finding equivalent resistances.

The problem is that interesting circuits *don’t* just involve resistors. They also contain things like inductors and capacitors. And *these* circuit elements don’t have that nice linear relationship between current and voltage. For capacitors, the relationship is between voltage and the *integral* of current. And for inductors, the relationship is between voltage and the *derivative* of current. Thus, a circuit involving a capacitor, a resistor, and an inductor, is going to be solved by an equation that involves the derivative of current, the current itself, and the integral of current. In other words, a circuit involving all three types of circuit elements is in general going to be solved by a second-order differential equation. And those are a mess.

The amazing thing is that if instead of treating current and voltage as real-valued functions, you treat them as complex-valued functions, what you find is that capacitors and inductors behave *exactly like resistors*. Voltage and current become related by a simple linear equation once more, with no derivatives or integrals involved. And the distinguishing characteristic of a capacitor or an inductor is that the constant of proportionality between voltage and current is an *imaginary number* instead of a real number. More specifically, a capacitor is a circuit element for which voltage is equal to a negative imaginary number times the current. An inductor is a circuit element for which voltage is equal to a positive imaginary number times the current. And a resistor is a circuit element for which voltage is equal to a positive *real* number times the current.

Suppose our voltage is described by a simple complex function: V_{0 }exp(iωt). Then we can describe the relationship between voltage and current for each of these circuit elements as follows:

Notice that now the equations for all three circuit components look just like resistors! Just with different constants of proportionality relating voltage to current. We can even redraw our original diagrams to make the point:

Fourier showed that *any function whatsoever* can be described as a sum of functions that look like Vo exp(iωt). And there’s a nice theorem called the *superposition* theorem that allows us to use this to solve any circuit, no matter what the voltage is.

So, let’s go back to our starting circuit.

What we can now do is just redraw it, with all capacitors and inductors substituted for resistors with imaginary resistances:

This may look just as complicated as before, but it’s actually much much simpler. We can use the rules for reducing resistor equations to solve an immensely simpler circuit:

Our circuit is now much simpler, but the original complexity had to be offloaded somewhere. As you can see, it’s been offloaded onto the complex (in both senses of the word) value of the resistance of our simplified circuit. But this type of complexity is *much* easier to deal with, because it’s just algebra, not calculus! To solve the current in our circuit now, we don’t need to solve a single differential equation, we just have to do some algebraic rearranging of terms. We’ll give the final equivalent resistance the name “Z”.

Now suppose that our original voltage was just the real part of V (that is, V_{0 }cos(ωt)). Then the current will also be the real part of I. And if our original voltage was just the imaginary part of V, then the current will be the imaginary part of I. And there we have it! We’ve solved our circuit without struggling over a single differential equation, simply by assuming that our quantities were complex numbers instead of real numbers.

This is one of my favorite examples of complex numbers playing an enormously important role in physics. It’s true that it’s a less clear-cut example of complex numbers being *necessary* to describe a physical phenomena, because we could have in principle done everything with purely real-valued functions by solving some differential equations, but it also highlights the way in which accepting a complex view of the world can simplify your life.