Fixed Points of Möbius Transformations

Möbius transformations are among the most elegant and fundamental functions in complex analysis, celebrated for their remarkable ability to map circles and lines to other circles and lines. This unique geometric property makes them indispensable tools in fields ranging from pure mathematics to theoretical physics. While these transformations describe a dynamic warping, stretching, and shifting of the complex plane, a profound question arises: in this whirlwind of motion, are there any points of stillness? These stationary points—the fixed points—are the secret anchors that govern the entire dynamical system.

This article delves into the world of these fixed points, revealing them to be far more than simple algebraic solutions. We will explore how their existence and nature provide a complete blueprint for the transformation's behavior. In the first chapter, ​​"Principles and Mechanisms,"​​ we will uncover the algebraic methods for finding fixed points and use them to create a powerful classification system for all Möbius transformations. In the second chapter, ​​"Applications and Interdisciplinary Connections,"​​ we will see how these theoretical anchors have profound implications, from defining motions in non-Euclidean geometry to revealing the deep algebraic structure that underpins them.

Imagine you have a magical, infinitely stretchable rubber sheet—the complex plane. A Möbius transformation is a special way of warping this sheet: it can stretch, shrink, rotate, and slide it around, but it always does so in a beautifully constrained way—circles and straight lines are always mapped to other circles or straight lines. After the introduction teased the elegance of these functions, our journey now takes us to their very heart. We will ask a simple but profound question: in this whirlwind of motion, are there any points of stillness?

When you spin a globe, two points remain perfectly still: the North and South Poles. These are the fixed points of the rotation. A Möbius transformation $T(z)$ also has such points of stillness, which we call ​​fixed points​​. They are the solutions to the simple-looking equation:

$T(z_0) = z_0$

Let's see what happens when we try to solve this. A typical Möbius transformation looks like $T(z) = \frac{az+b}{cz+d}$. So, our search for fixed points becomes a hunt for solutions to:

$z = \frac{az+b}{cz+d}$

A little bit of algebra transforms this into something very familiar. If we multiply both sides by $cz+d$ (assuming $z$ is not the point where the denominator is zero), we get:

$z(cz+d) = az+b$

$cz^2 + dz = az+b$

And by rearranging the terms, we arrive at a standard quadratic equation:

$cz^2 + (d-a)z - b = 0$

This is a wonderful moment! The search for stillness in a complex geometric transformation has boiled down to solving a high-school algebra problem. And we know that a quadratic equation can have, at most, two distinct solutions. This reveals a fundamental law: any Möbius transformation that isn't trivial has at most ​​two​​ fixed points in the finite complex plane. For instance, the transformation $g(z) = \frac{z-2}{z-1}$ has its fixed points at $z = 1+i$ and $z = 1-i$, the two roots of $z^2-2z+2=0$. Just like a spinning globe, most of these transformations have two poles that anchor their entire motion.

But what about the "point at infinity", which we include in our extended complex plane? A transformation can fix infinity as well. This happens if the transformation doesn't "grow" infinitely large. For $T(z) = \frac{az+b}{cz+d}$, the value $T(\infty)$ is the limit as $z$ gets enormous, which is simply $\frac{a}{c}$. For $\infty$ to be a fixed point, we need $T(\infty) = \infty$, which happens if and only if $c=0$. In that case, the transformation is just a linear map, $T(z) = \frac{a}{d}z + \frac{b}{d}$, and the quadratic equation for fixed points degenerates into a linear one.

This "at most two fixed points" rule is incredibly powerful. Let's play detective. Suppose a colleague comes to you and says, "I have found a miraculous Möbius transformation that has three distinct fixed points!". Should you be impressed?

Not at all! You know that the fixed points are the roots of the quadratic equation $cz^2 + (d-a)z - b = 0$. A non-zero quadratic equation cannot have three roots. The only way for this to be possible is if the equation is not a quadratic at all—if it is the trivial equation $0=0$. This requires all the coefficients to be zero:

$c = 0$ $d-a = 0 \implies d=a$ $b = 0$

What kind of transformation has $b=0$, $c=0$, and $a=d$? Let's plug these back into the formula: $T(z) = \frac{az+0}{0 \cdot z+a} = \frac{az}{a} = z$

It's the ​​identity transformation​​! The one that does nothing at all. Every point is a fixed point for the identity. So, the moment someone claims to have found three fixed points, you know their secret: their "miraculous" transformation is just the identity in disguise. This isn't a failure; it's a beautiful piece of rigidity. The structure of Möbius transformations is so tight that it forbids any non-trivial map from having more than two fixed points.

The number of fixed points gives us a powerful way to classify the geometry of these transformations.

• ​​Two Distinct Fixed Points:​​ This is the most common case, corresponding to the quadratic equation having two distinct roots. We will see that these transformations behave like flows from one pole to another, or rotations around them.

• ​​One Distinct Fixed Point:​​ What if the quadratic equation has a single, repeated root? This happens when the discriminant, $(d-a)^2 + 4bc$, is zero. These transformations are called ​​parabolic​​. The simplest example is a translation, $T(z)=z+1$. If you imagine sliding a vast sheet of paper one inch to the right, what point remains fixed? No finite point does! Every point moves. But the "point at infinity" can be thought of as remaining at infinity. A translation is a parabolic transformation whose single fixed point is at infinity. These transformations represent pure "shearing" or "sliding" motions.

The Fundamental Theorem of Algebra guarantees that every non-constant polynomial with complex coefficients has at least one complex root. This ensures that every non-identity Möbius transformation has at least one fixed point in the extended complex plane. There are no Möbius transformations that are completely un-anchored.

Fixed points are not just static markers; they are the dynamical heart of the transformation. They dictate what happens when you apply the map over and over again: $z_1 = T(z_0)$, $z_2 = T(z_1)$, $z_3=T(z_2)$, and so on.

To understand the behavior near a fixed point $z_0$, we look at the derivative, $T'(z_0)$. This complex number, sometimes called the ​​multiplier​​ $\lambda$, tells us how the transformation behaves locally. It acts like a scaling and rotation factor for tiny displacements around $z_0$. The magnitude $|\lambda| = |T'(z_0)|$ is key.

• If $|\lambda| \lt 1$, the transformation is a contraction near $z_0$. Each iteration pulls points closer to $z_0$. We call $z_0$ an ​​attractive​​ fixed point, or a sink. Any iterative process starting close enough will spiral into it.

• If $|\lambda| \gt 1$, the transformation is an expansion near $z_0$. Each iteration pushes points away. We call $z_0$ a ​​repelling​​ fixed point, or a source. It's a point of instability; a slight nudge away from it leads to a rapid departure.

• If $|\lambda| = 1$, points are neither systematically pulled in nor pushed away. They tend to orbit $z_0$. This is an ​​indifferent​​ or ​​neutral​​ fixed point.

This allows us to further classify the transformations with two fixed points:

• ​​Hyperbolic:​​ One fixed point is attractive, the other repelling. This happens when the multiplier $\lambda$ is a real number (and not 1). The transformation creates a "flow" on the Riemann sphere from the repelling source to the attractive sink.

• ​​Elliptic:​​ Both fixed points are indifferent. This happens when the multiplier has magnitude 1 (but isn't 1 itself, so $\lambda = e^{i\theta}$ for $\theta \neq 0$). The transformation corresponds to a pure rotation of the Riemann sphere around the axis connecting the two fixed points.

• ​​Loxodromic:​​ This is the general case, where one point is attractive and the other repelling, but the multiplier $\lambda$ is a complex number, not purely real. The motion is a spiral: points are simultaneously being pushed away from the repeller and towards the attractor, while also rotating around them. An airplane flying a rhumb line from one pole to another on a globe follows a loxodromic path.

Is there a way to capture the essence of these dynamics in a single, beautiful expression? The answer lies in one of the jewels of geometry: the ​​cross-ratio​​. For four distinct points, the cross-ratio is defined as $(z_1, z_2, z_3, z_4) = \frac{(z_1-z_3)(z_2-z_4)}{(z_1-z_4)(z_2-z_3)}$. The miracle of Möbius transformations is that they preserve this quantity.

Now, let's do something clever. Let $p$ and $q$ be the two fixed points of a transformation $T$. Let's compute the cross-ratio of a point $z$, its image $T(z)$, and the two fixed points: $(T(z), z, p, q)$. Because $T$ preserves cross-ratios, we have:

$(T(z), z, p, q) = (T(T(z)), T(z), T(p), T(q))$

But since $p$ and $q$ are fixed points, $T(p)=p$ and $T(q)=q$. So, we find that $(T(z), z, p, q) = (T(T(z)), T(z), p, q)$. This means the value of this cross-ratio is the same for the pair $(z, T(z))$ as it is for the pair $(T(z), T(T(z)))$, and so on. In fact, it's a constant, independent of the choice of $z$!

And what is this magical constant? It is nothing other than the multiplier, $\lambda = T'(p)$. This is a stunning unification: a purely geometric quantity (the cross-ratio) is precisely the number that governs the dynamics of the system (the multiplier).

This gives us the master equation for the dynamics of any non-parabolic Möbius transformation:

$(T(z), z, p, q) = \lambda$

By applying this repeatedly, we find an even more powerful relation for the $n$-th iterate of the transformation:

$(T^n(z), z, p, q) = \lambda^n$

This simple equation tells the whole story. If $|\lambda| \lt 1$ (the attractive case), then as $n \to \infty$, $\lambda^n \to 0$. For the cross-ratio to go to zero, the numerator $(T^n(z)-p)(z-q)$ must vanish, which implies that $T^n(z)$ must approach the fixed point $p$. The dynamics are encoded in a single, elegant formula.

To conclude, let's explore one final, beautiful piece of structure. What if we have two different Möbius transformations, $S$ and $T$, that ​​commute​​, meaning that applying $S$ then $T$ gives the same result as applying $T$ then $S$ ($S \circ T = T \circ S$). What does this "algebraic handshake" tell us about their "geometric layout"?

Let's assume $S$ has distinct fixed points $\{p_S, q_S\}$ and $T$ has distinct fixed points $\{p_T, q_T\}$, and all four points are different. Consider the point $T(p_S)$. Let's see what $S$ does to it:

$S(T(p_S)) = T(S(p_S))$ (by commutation) $S(T(p_S)) = T(p_S)$ (since $p_S$ is a fixed point of $S$)

This shows that $T(p_S)$ is a fixed point of $S$! So, $T(p_S)$ must be either $p_S$ or $q_S$. It cannot be $p_S$, because that would make $p_S$ a fixed point of $T$, but we assumed the fixed point sets were disjoint. Therefore, $T$ must map the fixed points of $S$ to each other: $T(p_S)=q_S$ and $T(q_S)=p_S$. By the same logic, $S$ must swap the fixed points of $T$.

Now we use the invariance of the cross-ratio. Let $C = (p_S, q_S, p_T, q_T)$. Applying the transformation $T$ to all four points shouldn't change the value of $C$:

$C = (T(p_S), T(q_S), T(p_T), T(q_T))$ $C = (q_S, p_S, p_T, q_T)$ (substituting what we found)

But how does $(q_S, p_S, p_T, q_T)$ relate to the original cross-ratio $C$? A quick look at the formula shows it is exactly $1/C$. So the commutation property forces the condition $C=1/C$, or $C^2 = 1$. The solutions are $C=1$ or $C=-1$. The value $C=1$ would imply some of the four distinct points were actually the same, which we ruled out.

This leaves only one possibility: the cross-ratio of the four fixed points must be $-1$. A set of four points with a cross-ratio of $-1$ is called a ​​harmonic quadruple​​. We have discovered a profound law: if two loxodromic transformations commute, their fixed points must form a harmonic quadruple. An algebraic property—commutation—enforces a precise, symmetric geometric arrangement. It is in these deep connections, where algebra and geometry dance in perfect synchrony, that the true beauty of Möbius transformations is revealed.

Now that we have grappled with the machinery of finding fixed points, you might be wondering, "What is all this for?" It is a fair question. In science, we are not merely stamp collectors, cataloging mathematical objects for their own sake. We seek to understand nature, and the tools we build must help us in that quest. The fixed points of Möbius transformations are not just idle curiosities; they are the lighthouses in the fog, the anchors in the storm, the very poles around which the dynamics of these transformations revolve. They are the key that unlocks a treasure chest of applications, revealing astonishingly deep connections between geometry, algebra, and even the fabric of spacetime. Let us embark on a journey to explore this new territory.

Imagine dropping a leaf into a swirling stream. While most of the water is in constant motion, there might be a still point in an eddy, a place where the leaf would simply spin in place. This is the essence of a fixed point. For Möbius transformations, these "still points" govern the entire flow.

Consider a hyperbolic transformation, which, as we've seen, has two fixed points. If you were to plot the paths of points under repeated applications of this transformation, you would not see chaos. Instead, a beautiful and orderly pattern emerges. Points flow from one fixed point (the repeller) to the other (the attractor). But what about the sets that are left unchanged by this flow? It turns out there is a whole family of circles, known as Apollonian circles, that are mapped perfectly onto themselves. Each of these circles is defined by the property that for any point $z$ on it, the ratio of its distances to the two fixed points, $z_1$ and $z_2$, is constant. The transformation simply slides points along these invariant circular tracks. The fixed points act as the geometric "poles" that define this entire system of circles.

This relationship is so fundamental that we can turn the problem on its head. Suppose you have two circles in the plane and you want to find a transformation that leaves both of them perfectly invariant. This sounds like a daunting design task. Yet, the principle of fixed points gives us an elegant solution. If a transformation leaves a circle invariant, its fixed points must be inverse points with respect to that circle. Therefore, the fixed points we seek must be inverse points with respect to both circles simultaneously! By solving this purely geometric condition, we can pinpoint the locations of the fixed points, and from them, construct the unique transformation with the desired property. The fixed points contain the complete blueprint of the transformation.

This idea of fixed points as "poles" becomes even more vivid when we consider elliptic transformations. These correspond to rotations of the Riemann sphere. And what is a rotation? It is a motion defined by an axis. The two points where the axis pierces the sphere are the points that do not move—they are the fixed points! So, an elliptic Möbius transformation is nothing more than a rotation of the complex plane (as a sphere) about the axis connecting its two fixed points. When you compose two such rotations, say, rotating the sphere first around the east-west axis and then around the north-south axis, you get a new, single rotation about some new, tilted axis. Finding the fixed points of the composite transformation is equivalent to finding the poles of this new rotation.

For centuries, Euclid's geometry was considered the only possible geometry. But in the 19th century, mathematicians discovered a new, logically consistent world: hyperbolic geometry. In this world, the "shortest distance" between two points is not a straight line but an arc of a circle, and the angles of a triangle always add up to less than 180 degrees. One of the most beautiful models of this strange new world is the Poincaré disk, the set of all complex numbers $z$ with $|z| \lt 1$.

What are the "rigid motions" in this world? How can we move figures around without distorting their hyperbolic shape or size? The answer, astonishingly, is a specific subset of Möbius transformations! Fixed points now take on a profound physical meaning. A transformation with a single fixed point inside the disk represents a pure rotation around that point in hyperbolic space. A transformation with two fixed points on the boundary of the disk represents a "translation"—a motion along the unique hyperbolic "line" (a circular arc) that connects those two boundary points. The location of the fixed points relative to the space tells us the very nature of the motion.

The picture becomes even more cinematic when we step up to three dimensions. One model for 3D hyperbolic space is the upper half-space of points $(x, y, t)$ where $t>0$. The boundary of this universe is the complex plane. The orientation-preserving isometries, or rigid motions, of this space are again given by Möbius transformations acting on the boundary plane. Here, a loxodromic transformation—one with two fixed points and a complex multiplier—is revealed to be a magnificent ​​screw motion​​. It consists of a translation along the hyperbolic line (a vertical axis or a semicircle) connecting its two fixed points, combined with a simultaneous rotation around that same axis. An object moving under such a transformation spirals along its path, like a corkscrew through space. What was once an abstract classification based on multipliers is now a tangible, dynamic, and beautiful motion. This is not just a mathematical curiosity; the geometry of hyperbolic space is a cornerstone of Einstein's theory of special relativity and plays roles in modern cosmological models.

So far, we have viewed our subject through the lens of a geometer. But an algebraist sees something different. For every Möbius transformation, there is a $2 \times 2$ matrix lurking in the shadows. The composition of transformations corresponds to the multiplication of matrices. This is where the story takes a turn towards the abstract, yet immensely powerful, world of group theory.

The entire geometric character of a transformation—is it parabolic, elliptic, or hyperbolic?—is encoded in a single number derived from its matrix: the trace, $\tau$. The transformation will have exactly one fixed point (parabolic) if and only if $\tau^2 = 4$. If $\tau^2 \neq 4$, it will have two distinct fixed points. This simple algebraic condition, $\tau^2=4$, is the sharp dividing line between entirely different geometric worlds! This isn't just a computational trick; it tells us that the classification of motions on the sphere is identical to the classification of conjugacy classes in the group of $2 \times 2$ matrices, $SL(2, \mathbb{C})$. The geometry and the algebra are two sides of the same coin.

This algebraic viewpoint provides us with incredibly powerful tools. Suppose you want to analyze a complicated transformation, like the composition of several others. The direct approach can be a nightmare of algebra. But there is a more elegant way, a principle known as ​​conjugation​​. If you have a transformation $T$ with a fixed point $z_0$, you can find a simpler transformation $h$ that moves $z_0$ to a more convenient location, like $\infty$. The original transformation viewed from this new perspective, given by $S = h \circ T \circ h^{-1}$, becomes much simpler—it's often just a simple scaling and translation, an affine map. We can then easily analyze $S$, solve our problem in this simpler world, and then use the inverse transformation $h^{-1}$ to bring the answer back to our original world. This is like translating a difficult problem into a language where the solution is obvious, and then translating the answer back. It is a cornerstone of modern mathematics and physics. In this sense, fixed points are not just static landmarks; they are handles we can grab to manipulate and simplify our problems.

The algebraic structure allows us to build and study new transformations by combining old ones, for instance by looking at expressions like $S \circ T \circ S^{-1} \circ T^{-1}$. These are not arbitrary combinations; they are related to fundamental group-theoretic objects like commutators, which measure the degree to which two operations fail to be interchangeable.

So, we have come full circle. The humble fixed point, the solution to the simple equation $f(z)=z$, has proven to be a concept of extraordinary depth and utility. It is the axis of a rotation, the source and sink of a flow, the invariant pole of a geometric construction. It is a portal to the non-Euclidean world of hyperbolic geometry and a key to decoding the algebraic structure of matrix groups. In the elegant dance of Möbius transformations, the fixed points are the silent choreographers, dictating every step and revealing the profound and beautiful unity of mathematics.