Subharmonic Functions

In mathematics and physics, many systems are described by a state of perfect equilibrium, captured by harmonic functions. However, countless phenomena exhibit a natural bias or one-sided pressure, deviating from this ideal balance. This article delves into the elegant world of ​​subharmonic functions​​, the mathematical tool designed to model precisely this kind of non-equilibrium behavior. We will bridge the gap between the intuitive idea of "upward curvature" and its rigorous mathematical consequences. The following chapters will first lay out the foundational principles and mechanisms, and then explore the far-reaching impact of these concepts across various fields. You will learn how subharmonicity is defined, why it leads to the celebrated maximum principle, and how it is applied in complex analysis, partial differential equations, and even large-scale geometric analysis.

Imagine a perfectly stretched, flat rubber sheet. This is our "harmonic" state, a state of perfect balance. Now, what if we start to deform it? What if, at every point, the surface has a tendency to curve upwards, like the inside of a bowl? This simple, intuitive idea is the key to unlocking the rich world of ​​subharmonic functions​​. While a harmonic function represents equilibrium, a subharmonic function captures a tendency, a pressure, a one-sided bias. It’s a concept that beautifully generalizes ideas of curvature and averages, with consequences that are as profound as they are elegant.

For a smooth, twice-differentiable function $u$, which we can visualize as a surface, its deviation from being harmonic is measured by the ​​Laplacian​​, $\Delta u$. In two dimensions, this is $\Delta u = \frac{\partial^2 u}{\partial x^2} + \frac{\partial^2 u}{\partial y^2}$. If $u$ is harmonic, $\Delta u = 0$. For a subharmonic function, we define the condition as $\Delta u \ge 0$. This single inequality is a treasure trove of information. A non-negative Laplacian means the function has an inherent "upward" curvature, averaged over all directions.

Think about a taut membrane again. If forces are pulling it gently downwards from underneath, any point on the membrane will naturally be lower than the average height of the circle of points surrounding it. This is the famous ​​sub-mean value property​​: for a subharmonic function, its value at the center of a ball is less than or equal to the average of its values on the ball's surface.

This isn't just an abstract inequality; we can see it in action. Consider the simple function $u(x) = |x|^2$ over $\mathbb{R}^n$. This is the very shape of a bowl. Its Laplacian is $\Delta u = 2n$, which is strictly positive for $n \ge 1$, so it's subharmonic. If we calculate the average of $u$ on a sphere of radius $r$ centered at $x_0$, we find it is exactly $|x_0|^2 + r^2$. The "mean value gap" is precisely $r^2$, a positive number, confirming that the center value $|x_0|^2$ is strictly less than the average.

But what if a function isn't smooth? What about a function with a "kink," like $u(x)=|x|$, or a singularity, like $u(x) = \ln|x|$ in the plane? We can't compute a Laplacian at the point of non-differentiability. Here, mathematics performs a clever trick. Instead of inspecting the function at every point, we define its Laplacian "in the sense of distributions." The idea is to understand the function's curvature by how it interacts with infinitely smooth, localized "bump" functions $\phi$ (called test functions). A function $u$ is subharmonic if, for every non-negative test function $\phi$, we have $\int u (\Delta \phi) \, dx \ge 0$. This definition beautifully sidesteps the need for derivatives of $u$ itself, allowing us to analyze a much broader class of functions.

The results can be stunning. For the function $u(x) = \ln|x|$ in the plane, which is harmonic everywhere except the origin, this generalized method reveals that its distributional Laplacian isn't zero; it is $\Delta u = 2\pi\delta_0$. This is the ​​Dirac delta function​​—a distribution that is zero everywhere except for an infinitely sharp "spike" of positive measure concentrated entirely at the origin. So, $\ln|x|$ is indeed subharmonic. Its "upward curvature" is entirely focused at a single point!

The single most important consequence of the sub-mean value property is the ​​maximum principle​​. The logic is almost poetic: if the value at every point is no greater than the average of its neighbors, then how could any point be a strict peak, towering above all its neighbors? It's impossible.

More formally, the ​​strong maximum principle​​ states that a non-constant subharmonic function cannot attain a maximum value in the interior of its domain. If it has a maximum, that maximum must lie on the boundary. For a continuous subharmonic function on a bounded domain like a disk, this means $\sup_{\text{domain}} u = \sup_{\text{boundary}} u$. For $u(x) = \ln|x|$ on a disk of radius $R$, its minimum is at the center ($-\infty$), and its values increase outwards, reaching their maximum of $\ln R$ only on the boundary circle. Subharmonic functions are always "pushing" their highest values out to the edges.

This principle has a curious asymmetry. It forbids interior maxima, but it says nothing about interior minima. The function $u(x) = |x|^2$ is a perfect example: it is subharmonic, but it has a very clear interior minimum at the origin.

What if the domain has no boundary, like the entire plane or a complete manifold? In this case, the maximum principle takes on an even more powerful form, often called a ​​Liouville-type theorem​​. On a complete manifold with non-negative Ricci curvature (a geometric condition that generalizes "flatness"), any subharmonic function that is bounded above must be constant. There's simply "no room" for it to vary. This notion can be extended through the remarkable ​​Omori-Yau maximum principle​​, which tells us that even if a bounded-above subharmonic function on such a manifold doesn't attain its maximum, we can still find a sequence of points "approaching the top" where the function becomes arbitrarily flat and its Laplacian approaches zero. This powerful tool reveals another gem: if a function is "strictly" subharmonic everywhere (e.g., $\Delta u \ge c > 0$), it cannot be bounded above at all; it must grow to infinity. The constant upward pressure guarantees it will never level off.

The maximum principle tells us subharmonic functions are wonderfully "well-behaved." Can we use this behavior to solve problems? One of the fundamental problems in physics and mathematics is the ​​Dirichlet problem​​: given a set of values on a boundary, find a function inside that is harmonic ($\Delta u = 0$) and matches those boundary values. This models everything from steady-state heat distribution to electrostatic potentials.

The ​​Perron method​​ provides an astonishingly elegant way to construct the solution. Instead of trying to find the harmonic function directly, we build it from a family of subharmonic functions. Imagine we have our boundary values specified by a function $g$. Now, consider the set of all subharmonic functions that live inside the domain and whose values stay at or below $g$ on the boundary. Each of these functions is a valid "lower approximation" to the solution we seek.

Now for the magic step: we define a new function, $U(z)$, by taking the pointwise "ceiling" (the supremum) of this entire family of lower approximations.

One might expect $U(z)$ to be a complicated, perhaps even subharmonic, function. But the beautiful theorem of Perron states that this function $U(z)$ is the one and only ​​harmonic​​ function that solves the Dirichlet problem. It’s as if by packing together all possible well-behaved "floors," we perfectly construct the smooth, equilibrium ceiling.

Let's return to the idea of averaging. Let $M(r)$ be the average of a subharmonic function $u$ on a circle of radius $r$. We already know that $u$'s value at the center is less than $M(r)$. More than that, the function $M(r)$ itself has a beautiful property: it is a ​​convex function of $\log r$​​.

This might sound technical, but its meaning is profound. Convexity is a statement about straight lines. If we plot $M(r)$ versus $\log r$, the graph of the function will always lie below the straight line segment connecting any two of its points. This places a powerful constraint on the growth of the average values. For example, if we know the average value on a circle of radius 1 and on a circle of radius 25, we can immediately establish a firm upper bound for the average value on any circle in between, like the one with radius 5.

This interplay between the local property of a function ($\Delta u \ge 0$) and its global or averaged behavior is a recurring theme. It shows how a simple, local rule of "upward curvature" echoes through the entire domain, constraining the function in ways that are both surprisingly rigid and exquisitely beautiful. From the gentle curve of a bowl to the abstract geometry of manifolds, the principles of subharmonicity provide a unified language to describe states of non-equilibrium and the elegant laws that govern them.

You might be thinking, "This is all very elegant, but what is it for?" That is a fair and essential question. The beauty of a concept in science is often revealed not just in its internal logic, but in how it reaches out and touches other ideas, solving problems and providing new ways of seeing the world. Subharmonic functions are a prime example of this. Having explored their fundamental principles, we're now ready for a journey to see them in action, a journey that will take us from practical calculations to the vast, curved expanses of the cosmos.

Let’s start with the most immediate and striking consequence: the Maximum Principle. Imagine you have a thin metal plate, and you're heating its edges. The temperature inside the plate will eventually reach a steady state. A physicist might tell you that, in the absence of any heat sources within the plate itself, the hottest point will always be somewhere on the edge, never in the middle. This physical intuition is perfectly captured by mathematics. The temperature function in this scenario is harmonic, a special case of a subharmonic function.

The principle holds much more generally. As we've learned, if a function $f(z)$ is analytic, then its modulus $|f(z)|$ is subharmonic. This is an enormous source of examples! It gives us a powerful tool for a task that is often very difficult: finding the maximum value of a function. Suppose we are handed a complicated-looking function, like $u(z) = |z^2 - 1/4|$, and asked to find its largest value within a given region, say, a square or a disk. The brute-force approach of checking points everywhere is daunting. But the Maximum Principle tells us not to bother with the interior! We know, with mathematical certainty, that the peak value must lie on the boundary. Our search is reduced from an infinite sea of possibilities in a two-dimensional area to a one-dimensional line. The same logic applies even if the region is more complex, like a donut-shaped annulus, where the principle guarantees the maximum must occur on either the inner or the outer circular boundary.

This principle isn't just for finding the single highest peak. It gives us a way to control or estimate the function's behavior everywhere. The sub-mean-value property, which lies at the heart of subharmonicity, states that the value at the center of any disk is no greater than the average of the values on its boundary circle. This has profound implications. If we have a subharmonic function $u(z)$ inside the unit disk and we know that on the boundary circle its value is limited—say, by another function like $|\operatorname{Re}(z^n)|$—we can place a strict, sharp limit on the value $u(0)$ at the very center. By simply averaging the bounding function over the circle, we discover that $u(0)$ can be no larger than $2/\pi$, a beautiful and rather unexpected constant that emerges from the geometry of the circle and the cosine function. This is a tool of immense power: knowledge on the boundary translates directly into control on the inside.

The Maximum Principle tells us that a subharmonic function's maxima are on the boundary, but it doesn’t quite tell us why. What is the "source" of this behavior? For smooth functions, the answer lies with the Laplacian operator, $\Delta = \frac{\partial^2}{\partial x^2} + \frac{\partial^2}{\partial y^2}$. A function $u$ is subharmonic if its Laplacian is non-negative, $\Delta u \ge 0$.

Think back to our heated plate. If the temperature is harmonic, $\Delta u = 0$, meaning heat flows in and out of any small region at an equal rate. But if we place tiny heat sources inside the plate, then more heat flows out than in, and we find $\Delta u > 0$. The Laplacian, then, is a measure of the density of "sources."

The Riesz Representation Theorem brilliantly generalizes this intuition to all subharmonic functions, even those that aren't smooth enough to have a Laplacian in the usual sense. It tells us that for any subharmonic function $u$, there is a unique recipe of sources—a positive measure $\mu_u$—such that $\Delta u = 2\pi \mu_u$. This "Riesz measure" tells us exactly where the subharmonicity is coming from and how strong it is.

For a simple function like $u(z) = |z - 1/2|^2$, the Laplacian is a constant, $\Delta u = 4$. This means the Riesz measure is spread perfectly evenly, like a fine dust, over the entire plane. To find the total "source strength" inside the unit disk, we simply multiply this constant density by the disk's area.

But the true magic appears with more structured functions. Consider taking the maximum of two subharmonic functions, like $u(z) = \max(\log|z^2-i|, \log|z^2+i|)$. The logarithm of the modulus of an analytic function is a classic subharmonic function whose Riesz measure consists of point-like sources at the function's zeros, just like point charges in electrostatics. When we take the maximum of two such functions, we are essentially building a new landscape by taking the upper envelope of two existing ones. The Riesz measure of this new function—its "source" distribution—is fascinating. It includes not only the original point charges but also a new source that is distributed precisely along the "crease" or "ridge" where one function takes over from the other. This gives us a startlingly beautiful geometric picture: the act of taking a maximum creates a new source of subharmonicity right along the seam.

The concept of subharmonicity is not confined to complex analysis. It is a central character in the world of Partial Differential Equations (PDEs). The condition $\Delta u \ge 0$ is a fundamental type of differential inequality. An interesting question one can ask is how to build new subharmonic functions from existing ones. For instance, if you start with a harmonic function $u$ (where $\Delta u = 0$), can you create a subharmonic function by combining $u$ and its derivatives? Consider the function $g = |\nabla u|^2 + k u^2$. A careful calculation of its Laplacian reveals that $g$ is guaranteed to be subharmonic for any harmonic function $u$ if and only if the constant $k$ is non-negative. This demonstrates a deep structural relationship, governed by the very rules of differentiation, between the flatness of the harmonic world and the curvature of the subharmonic one.

Another powerful idea is that of approximation. A subharmonic function can have bumps and kinks. We might ask, what is the smoothest possible function that stays entirely above our bumpy one? This "smoothest" upper bound is a harmonic function, called the least harmonic majorant. It's like stretching a perfectly flat sheet just over the top of a rugged landscape. Remarkably, potential theory gives us a way to construct this majorant. For a subharmonic function $v(z)$ on the unit disk, like $v(z) = \max(\text{Re}(z), \text{Im}(z))$, the value of its least harmonic majorant at the center of the disk is simply the average of the values of $v(z)$ on the boundary circle. This provides a powerful connection: the seemingly chaotic behavior of a subharmonic function can be "smoothed out" into a harmonic one in a canonical way, and the link between them is the simple act of averaging over a boundary.

So far, we have been living in the flat world of the Euclidean plane. But what happens if our universe itself is curved? What happens to subharmonic functions on the surface of a sphere, or a saddle-shaped surface that extends to infinity? This is where subharmonic functions enter the grand stage of geometric analysis. Here, the properties of functions are inextricably linked to the geometry of the space they inhabit.

A profound result, known as a Liouville-type theorem, states that on certain kinds of infinite, curved spaces, any non-negative subharmonic function must be a constant. Imagine a landscape with no dips that sits above sea level; if the geometry of the world is right, this landscape must be perfectly flat! Whether this astonishing rigidity holds depends on the curvature of the space—specifically, how fast its volume grows as you move away from a point. On a manifold that expands at a modest rate, there is "no room" for a subharmonic function to grow, and it is forced to be constant. If the space flares out too quickly, however, non-constant functions can exist. This is a stunning link between analysis (the behavior of $\Delta u$) and global geometry (the volume growth of the universe).

Finally, what becomes of the Maximum Principle in an infinite world that has no boundary? The Omori-Yau Maximum Principle provides the answer. It says that on a complete, non-compact manifold whose curvature doesn't get too wildly negative, any subharmonic function that is bounded above cannot achieve its supremum at some isolated peak in the interior and then fade away. Instead, its supremum is reflected in its behavior "at infinity." The function must get arbitrarily close to its maximum value as one travels ever farther out into the space. The geometry of the space acts as a global leash, preventing the function from hiding its peak in a finite region.

From a simple rule about the hottest point on a plate, we have journeyed to the structure of PDEs and the relationship between curvature and analysis on a cosmic scale. Subharmonic functions, born from the elegant world of complex numbers, prove to be a unifying thread, weaving together seemingly disparate fields and revealing, at every turn, the deep and inherent beauty of mathematical thought.