Clairaut's Theorem

In mathematics, as in life, the order of operations can be critical. While putting on shoes before socks yields a nonsensical result, other sequences are freely interchangeable. This raises a fundamental question in multivariable calculus: when we measure a function's rate of change in one direction, and then measure how that rate of change itself changes in another direction, does the order of our measurements matter? This inquiry into the "commutativity" of partial differentiation leads to one of the most elegant and surprisingly powerful results in mathematical analysis: Clairaut's Theorem. This article addresses the knowledge gap between simply knowing the rule and understanding its profound implications. You will learn the core principle of the theorem and the crucial conditions for its validity, before exploring how this seemingly simple mathematical symmetry unlocks deep connections in physics, simplifying the study of energy, thermodynamics, and fluid dynamics. We begin by examining the principles and mechanisms that govern this dance of derivatives.

Imagine you're getting dressed in the morning. You put on your socks, then you put on your shoes. The order is non-negotiable. Trying to do it the other way around leads to a rather comical and entirely unsuccessful result. Some operations in life, and in mathematics, are like this—they do not ​​commute​​. The order matters. But what about putting on your left sock and your right sock? The order is irrelevant; the outcome is the same. These operations ​​commute​​.

It turns out that one of the fundamental operations in calculus, taking a derivative, often behaves like putting on your socks. When we are in the realm of functions of multiple variables, say a function $f(x, y)$ that describes a hilly landscape, we can ask about its slope in different directions. The partial derivative with respect to $x$, written as $\frac{\partial f}{\partial x}$ or $f_x$, tells us the slope as we walk along the $x$-axis. Similarly, $\frac{\partial f}{\partial y}$ or $f_y$ tells us the slope along the $y$-axis.

But what if we take two derivatives? What if we first find the slope in the $x$-direction ($f_x$), and then ask how that slope changes as we move a tiny step in the $y$-direction? This gives us the mixed partial derivative $\frac{\partial}{\partial y}\left(\frac{\partial f}{\partial x}\right)$, which we denote as $f_{xy}$. Or, we could do it the other way around: find the slope in the $y$-direction ($f_y$), and then ask how it changes as we move in the $x$-direction. This is $\frac{\partial}{\partial x}\left(\frac{\partial f}{\partial y}\right)$, or $f_{yx}$.

The profound question is: should these two values be the same? Are we talking about shoes and socks, or left and right socks? On an intuitive level, both $f_{xy}$ and $f_{yx}$ seem to be measuring the same thing: the "twist" or "cross-curvature" of the surface at a point. It seems plausible that they should be equal. This very idea is captured by a beautiful and powerful result known as ​​Clairaut's Theorem​​.

Clairaut's Theorem states that for a "sufficiently well-behaved" function, the order of mixed partial differentiation does not matter. That is:

What does "sufficiently well-behaved" mean? We'll get to the fine print in a moment, but for now, let’s get a feel for this principle by watching it in action. Let's take a tour through a gallery of familiar functions and see how they dance to Clairaut's tune.

For the vast majority of functions you encounter in science and engineering—polynomials, exponentials, trigonometric functions, and their combinations—this property holds true without any fuss. Take a simple polynomial like $f(x, y) = 3x^4y^2 - 5x^2y^3 + 2y$. If you meticulously calculate $f_{xy}$ and then $f_{yx}$, you will find they are identical. The same elegant equality appears when we examine functions involving logarithms like $f(x, y) = y \ln(x)$, combinations of exponentials and trigonometric functions like $h(x, y) = x^2 e^y \sin y$, or composite functions like $f(x, y) = \cos(x^2 + y^2)$. Even rational functions, like $h(u, v) = \frac{u-v}{u+v}$ where the denominator is non-zero, obey this rule. It's not just a coincidence for specific functions; it's a deep property of their structure, as can be seen by proving it for a general form like $f(x, y) = (ax + by)^n$.

This symmetry isn't limited to just two derivatives. What if we differentiate three times, or four? For our well-behaved functions, the order never matters. If you take a function like $h(s, t) = s^3 t^3$ and calculate the third-order derivatives $h_{sst}$ (differentiating with respect to $s$, then $s$ again, then $t$) and $h_{sts}$ (differentiating $s$, then $t$, then $s$), you will find they are precisely the same. The operators of partial differentiation, for a vast class of functions, commute freely.

By now, you might be tempted to think this is a universal law of mathematics. But it is not a law; it is a theorem. And a theorem has conditions. The magic ingredient, the property that makes this whole beautiful symmetry work, is ​​continuity​​. Clairaut's Theorem only holds if the mixed second partial derivatives themselves exist and are ​​continuous​​ around the point of interest.

So, what happens if this condition is not met? Can we find a function where the symmetry breaks? Indeed, we can. Consider this rather peculiar function:

This function is continuous everywhere, even at the origin. You can even calculate its first partial derivatives, $f_x$ and $f_y$, everywhere. But at the origin, something strange happens. The second partial derivatives are not continuous. They exist, but they jump around erratically as you approach the origin from different directions.

If we ignore this warning sign and proceed, we are in for a surprise. By carefully applying the limit definition of the derivative at the origin, we can compute the mixed partials. The result is shocking:

But if we compute in the other order, we find:

They are not equal! The commuting property has vanished. This "counterexample" is a wonderful illustration of the scientific method. The exception doesn't just break the rule; it illuminates it. It shows us precisely which pillar—the continuity of the second derivatives—the entire structure rests upon. For most physical systems described by smooth functions, this condition is naturally met, which is why we can so often rely on Clairaut's a priori. For instance, if a system is described by a smooth implicit equation, the ​​Implicit Function Theorem​​ guarantees that the solution is also smooth, and therefore Clairaut's Theorem must hold.

Why is this theorem so important? Is it just a mathematical curiosity? Far from it. This symmetry is a cornerstone of many fields of science and a powerful tool in a mathematician's arsenal.

First, it simplifies our lives enormously. Suppose you are faced with a function defined by a complicated integral, like $f(x,y) = \int_0^x \int_0^y g(s,t) ds dt$. If you need to calculate $f_{xy}$, one order of differentiation might lead you down a very thorny path. But Clairaut's Theorem gives you the freedom to choose the easy way. By swapping the order of differentiation, you can often combine it with the ​​Fundamental Theorem of Calculus​​ to make the problem almost trivial. You get to choose the path of least resistance.

More profoundly, this symmetry is woven into the fabric of physics, particularly in ​​thermodynamics​​. Physical quantities like internal energy ($U$), enthalpy ($H$), and Gibbs free energy ($G$) are ​​state functions​​. This means their value depends only on the current state of the system (e.g., its temperature and pressure), not on the path taken to get there. Mathematically, this means their differentials are ​​exact differentials​​. An exact differential of a function $F(x,y)$, written $dF = M(x,y) dx + N(x,y) dy$, must satisfy the condition $\frac{\partial M}{\partial y} = \frac{\partial N}{\partial x}$. But since $M = \frac{\partial F}{\partial x}$ and $N = \frac{\partial F}{\partial y}$, this condition is nothing more than $F_{yx} = F_{xy}$—it's Clairaut's Theorem in disguise! The famous ​​Maxwell Relations​​ in thermodynamics, which connect seemingly unrelated properties of a substance, are all direct consequences of applying Clairaut's theorem to the thermodynamic potentials.

So, this simple idea—that for smooth functions, the order of differentiation doesn't matter—is far from a mere technical detail. It reveals a deep symmetry in the mathematical descriptions of our world. It gives us practical shortcuts, underpins fundamental laws of physics, and serves as a constant reminder that in the smooth, continuous landscape of functions that model our reality, the path we take to measure a change in change often doesn't matter at all. The destination is the same.

You might be thinking, "Alright, I see that for nice, smooth functions, the order in which I take partial derivatives doesn't matter. It's a neat mathematical trick. But what is it good for?" That is a fair and excellent question. The answer, it turns out, is that this simple-looking rule, Clairaut's Theorem, is not just a mathematical curiosity. It is the secret key that unlocks some of the most profound and useful concepts in all of physics and engineering. It is the cornerstone for the idea of a ​​potential​​, and through that, it sheds light on the hidden unity and symmetries of the physical world.

Imagine you are hiking in a mountainous national park. The force of gravity pulls you downwards. That force depends on your location—it might be slightly different at different latitudes or altitudes. But we know something wonderful about gravity: the total work you do against it to get from point A to point B depends only on the change in your altitude, not on the winding, zig-zagging path you took to get there. You could climb straight up a cliff face or take a long, gentle switchback; the change in your gravitational potential energy is exactly the same.

This is the essence of a ​​conservative force​​. The force field can be described as the gradient of a single scalar function, which we call the potential energy. Instead of having to keep track of a force vector at every point in space, we can just use a single, simpler "map" of the potential energy. This is an enormous simplification!

So, how do we know if a force field, or any vector field for that matter, is conservative? How do we know if a potential function even exists for it? Let's say we have a two-dimensional field described by components $M(x,y)$ and $N(x,y)$. If a potential function $f(x,y)$ exists, then we must have $M = \frac{\partial f}{\partial x}$ and $N = \frac{\partial f}{\partial y}$. Now, what happens if we differentiate $M$ with respect to $y$, and $N$ with respect to $x$?

$\frac{\partial M}{\partial y} = \frac{\partial}{\partial y} \left( \frac{\partial f}{\partial x} \right) = \frac{\partial^2 f}{\partial y \partial x}$ $\frac{\partial N}{\partial x} = \frac{\partial}{\partial x} \left( \frac{\partial f}{\partial y} \right) = \frac{\partial^2 f}{\partial x \partial y}$

If the potential function $f$ is "well-behaved" (meaning its second partial derivatives are continuous), then Clairaut's theorem tells us these two results must be equal. And thus, we have a test! To check if a field described by $(M, N)$ is conservative, we simply check if $\frac{\partial M}{\partial y} = \frac{\partial N}{\partial x}$. This famous test for ​​exact differential equations​​ is nothing more than Clairaut's theorem in a clever disguise. It's a quick check to see if a physicist's grand shortcut—the potential—is available for use.

Now, let's journey from the world of mechanics into the seemingly far-removed realm of thermodynamics—the study of heat, work, and energy. Here, we encounter quantities like temperature ($T$), pressure ($P$), volume ($V$), and entropy ($S$). We also encounter various forms of energy, such as the internal energy ($U$), enthalpy ($H$), and Gibbs free energy ($G$). These energies are "state functions," meaning their value depends only on the current state of the system (its temperature, pressure, etc.), not the history of how it got there.

Because they are well-behaved state functions, their differentials are exact. This means they act just like our potential functions from before, and therefore Clairaut's theorem must apply to them. And this is where the real magic happens.

Consider the enthalpy, $H$, which is a function of entropy $S$ and pressure $P$. The fundamental thermodynamic relation for its differential is:

$dH = T dS + V dP$

Just by looking at this, we can identify that $T = (\frac{\partial H}{\partial S})_P$ and $V = (\frac{\partial H}{\partial P})_S$. Now, let's apply Clairaut's theorem. We take the second partial derivatives of $H$, but in opposite orders:

$\frac{\partial}{\partial P} \left( \frac{\partial H}{\partial S} \right) = \frac{\partial}{\partial S} \left( \frac{\partial H}{\partial P} \right)$

Substituting what we know about the first derivatives gives us a stunning result:

$\left( \frac{\partial T}{\partial P} \right)_S = \left( \frac{\partial V}{\partial S} \right)_P$

This is one of the ​​Maxwell relations​​. Stop and think about what this says. It claims that the rate at which temperature changes as you increase pressure (while keeping entropy constant) is exactly equal to the rate at which volume changes as you increase entropy (while keeping pressure constant). Who would have guessed? There is no obvious physical reason why these two completely different processes should be so intimately linked. Yet, Clairaut's theorem shows us that this surprising connection is a necessary consequence of the existence of a state function called enthalpy. These relations are incredibly powerful because they allow experimentalists to determine quantities that are very difficult to measure (like the change in entropy) by measuring things that are much easier (like changes in temperature, pressure, and volume). This same principle extends to other coupled systems, revealing hidden relationships between the magnetic, thermal, and elastic properties of materials.

Let's turn our attention to the elegant and complex motion of fluids. To describe a flowing river, we can assign a velocity vector $\mathbf{v}$ to every point. Often, it's convenient to describe this flow using a ​​velocity potential​​ $\phi$, where the velocity field is simply the gradient of this scalar potential: $\mathbf{v} = \nabla \phi$. This is a huge simplification, reducing a three-component vector field to a single scalar function.

However, this only works for a special type of flow known as ​​irrotational flow​​—a smooth, laminar flow without any little eddies or whirlpools. The mathematical measure of the "whirlpool-ness" or local rotation in a fluid is called the ​​vorticity​​, defined as the curl of the velocity field, $\boldsymbol{\omega} = \nabla \times \mathbf{v}$.

So, what is the vorticity of a flow that comes from a potential? It's the curl of a gradient: $\boldsymbol{\omega} = \nabla \times (\nabla \phi)$. And what is the curl of a gradient? A quick check of the vector calculus identity reveals that it is always zero! But why? If you write out the components of the curl, you'll find they are all of the form $\frac{\partial^2\phi}{\partial x \partial y} - \frac{\partial^2\phi}{\partial y \partial x}$. Thanks to our friend Clairaut's theorem, these terms vanish for any smooth potential $\phi$. So, the fact that potential flow is irrotational is a direct consequence of the symmetry of mixed partial derivatives. It provides a deep link between the geometrical nature of the flow and the existence of a simplifying potential.

All these applications—conservative forces, exact equations, Maxwell relations, irrotational flow—point to the same underlying structure. Mathematicians have developed a beautiful and powerful language to describe this structure: the language of ​​differential forms​​.

In this language, a function $f$ is a "0-form." Its differential, $df = \frac{\partial f}{\partial x} dx + \frac{\partial f}{\partial y} dy$, is a "1-form." There is an operation, the "exterior derivative" $d$, that turns a $k$-form into a $(k+1)$-form. A 1-form $\omega$ is called ​​exact​​ if it's the derivative of some 0-form, i.e., $\omega = df$. A 1-form is called ​​closed​​ if its own derivative is zero, i.e., $d\omega = 0$.

Clairaut's theorem finds its most elegant expression here. When you apply the exterior derivative twice to a function $f$, you are essentially calculating the mixed partials. The profound statement that ​​every exact form is closed​​, written simply as $d(df) = 0$, is a cornerstone of differential geometry. And when you unpack its meaning in local coordinates, it is precisely the statement that $\frac{\partial^2 f}{\partial y \partial x} - \frac{\partial^2 f}{\partial x \partial y} = 0$. What we saw as a property of calculus is, from a higher vantage point, a fundamental principle of geometry itself.

Is this symmetry of differentiation, then, an absolute and unbreakable law of the universe? Not quite. The theorem comes with a crucial condition: the second partial derivatives must be continuous. While the functions that describe physical laws in our universe are typically wonderfully smooth, mathematicians can construct "pathological" functions where this condition fails. For such functions, which may not be differentiable or have discontinuous second derivatives at a point, the order of differentiation can indeed matter, and the Hessian matrix can fail to be symmetric.

Far from being a problem, these counterexamples are illuminating. They remind us that the elegant connections and labor-saving shortcuts we've explored are not just happy accidents. They are a direct consequence of the smoothness and regularity of the underlying potentials that govern the physical world. The universe, it seems, has a preference for functions that are well-behaved, and because of that, it presents us with the beautiful and surprising symmetries revealed by Clairaut's theorem.