Holomorphic function

Intuition [Beginner]

A holomorphic function is a function of a complex variable $z=x+iy$ that is differentiable in the complex sense: the limit defining its derivative exists and is independent of the direction in which you approach the point. This single requirement — direction-independent differentiability — is far stronger than real differentiability and forces a function's behaviour to be remarkably rigid.

Holomorphic functions are essentially "infinitely smooth" automatically: any holomorphic function is analytic (locally given by a convergent power series), differentiable infinitely many times, and the value of a holomorphic function on a small disk determines its value on a much larger region. A function that's holomorphic on a connected open set is determined globally by its behaviour on any tiny neighbourhood, by analytic continuation.

This rigidity is the source of complex analysis's striking theorems: Cauchy's theorem (loop integrals vanish), the residue theorem, the maximum modulus principle, and Liouville's theorem (a bounded entire function is constant).

Visual [Beginner]

A holomorphic function preserves angles locally — small squares map to small squares, rotated and scaled. This conformal property visualises the rigidity of holomorphic functions.

Worked example [Beginner]

The function $f(z)=z^2$ is holomorphic everywhere on $\mathbb{C}$. To check this directly: take a small complex number $h$ and form the difference quotient $((z+h{)}^2-z^2)/h=2z+h$. As $h$ shrinks to zero (along any direction), this approaches $2z$, so the limit exists and is direction-independent. The derivative is $f^{\prime }(z)=2z$.

By contrast, the conjugation function $g(z)=\bar{z}$ (sending $x+iy$ to $x-iy$) is not holomorphic anywhere. Approaching from the real direction $h\to 0$ along the $x$-axis gives a difference quotient of $1$, but approaching along the imaginary axis gives $-1$. The two directional limits disagree, so the complex derivative does not exist. The conjugation map is smooth as a real two-variable function but fails the rigid one-direction-fits-all complex condition.

The contrast — $z^2$ holomorphic, $\bar{z}$ not — captures the essence: holomorphicity is a strong rigidity, far beyond ordinary smoothness.

Check your understanding [Beginner]

Formal definition [Intermediate+]

Let $U\subseteq \mathbb{C}$ be an open subset. A function $f:U\to \mathbb{C}$ is holomorphic at $z_0\in U$ if the limit

$$f^{\prime }(z_0):=\underset{h\to 0}{\lim }\frac{f(z_0+h)-f(z_0)}{h}$$

exists, where $h\in \mathbb{C}\setminus \{0\}$ approaches $0$ in $\mathbb{C}$ (any direction). $f$ is holomorphic on $U$ if it is holomorphic at every point of $U$.

Equivalent characterisation (Cauchy-Riemann). Writing $f(z)=u(x,y)+iv(x,y)$ with $z=x+iy$, $f$ is holomorphic on $U$ iff $u,v$ are continuously differentiable on $U$ and satisfy

$$\frac{\partial u}{\partial x}=\frac{\partial v}{\partial y},\frac{\partial u}{\partial y}=-\frac{\partial v}{\partial x}.$$

In compact form: $\bar{\partial }f:=\frac{1}{2}\left(\frac{\partial f}{\partial x}+i\frac{\partial f}{\partial y}\right)=0$. Holomorphic functions are exactly the kernel of the Cauchy-Riemann operator $\bar{\partial }$.

Equivalent characterisation (analytic). $f$ is holomorphic on $U$ iff for every $z_0\in U$, there is a disk $D(z_0,r)\subseteq U$ on which $f$ has a convergent power series expansion:

$$f(z)=\sum _{n=0}^{\infty }{a}_n(z-z_0{)}^n,|z-z_0|<r,$$

where $a_n=f^{(n)}(z_0)/n!$. This is the analyticity of holomorphic functions.

Properties:

• Closure under operations: sum, product, composition of holomorphic functions are holomorphic; quotient is holomorphic where the denominator is nonzero.
• Cauchy's theorem: for $f$ holomorphic on a simply connected domain $U$ and $\gamma$ a closed loop in $U$, ${\oint }_{\gamma }f(z)dz=0$.
• Cauchy integral formula: for $f$ holomorphic on $U$ and a loop $\gamma$ around $z_0$ contained in $U$,$$f(z_0)=\frac{1}{2\pi i}{\oint }_{\gamma }\frac{f(z)}{z-z_0}dz.$$
• Maximum modulus principle: a non-constant holomorphic function does not achieve its maximum modulus in the interior of its domain.
• Liouville's theorem: a bounded entire function (holomorphic on all of $\mathbb{C}$) is constant.
• Identity theorem: two holomorphic functions on a connected open set that agree on a set with a limit point in the set must be equal.

Notation. A holomorphic function is sometimes called an analytic function or regular function. The space of holomorphic functions on $U$ is $\mathcal{O}(U)$ (or $H(U)$); this forms a sheaf of rings on $\mathbb{C}$ (and more generally on Riemann surfaces).

Key theorem with proof [Intermediate+]

Theorem (Cauchy's integral theorem). Let $f$ be holomorphic on a simply connected open set $U\subseteq \mathbb{C}$, and let $\gamma$ be a piecewise-smooth closed curve in $U$. Then

$${\oint }_{\gamma }f(z)dz=0.$$

Proof sketch. Write $f(z)dz=(u+iv)(dx+idy)=(udx-vdy)+i(vdx+udy)$. Apply Green's theorem to each real-valued line integral over a region $D$ bounded by $\gamma$:

$${\oint }_{\gamma }(udx-vdy)={\iint }_D\left(-\frac{\partial v}{\partial x}-\frac{\partial u}{\partial y}\right)dxdy=0$$

by Cauchy-Riemann, and similarly for the imaginary part. So ${\oint }_{\gamma }fdz=0$. $\square$

This Green's-theorem argument requires $u,v$ to have continuous partial derivatives, which is automatic for holomorphic $f$ (in fact, holomorphic functions are infinitely differentiable). A more refined proof (Goursat's theorem) avoids this hypothesis.

The integral theorem is the foundation of complex analysis: the Cauchy integral formula, residue theorem, contour integration, and analytic continuation all follow.

Bridge. The construction here builds toward 06.03.01 (riemann surface), where the same data is developed in the next layer of the strand. The defining pattern appears again in those units in a sharpened form, where the local data is glued or quotiented. Putting these together, the foundational insight is that the data of this unit gives the structural signature that the rest of the strand reads off.

Exercises [Intermediate+]

Lean formalization [Intermediate+]

lean_status: partial — Mathlib has Complex.HolomorphicAt, Complex.HolomorphicOn, the Cauchy integral formula, and Liouville's theorem.

[object Promise]

Advanced results [Master]

Open mapping theorem. A non-constant holomorphic function on a connected open set is an open map — it sends open sets to open sets.

Schwarz lemma. If $f:D\to D$ is a holomorphic map from the unit disk to itself with $f(0)=0$, then $|f(z)|\le |z|$ and $|f^{\prime }(0)|\le 1$, with equality iff $f$ is a rotation.

Riemann mapping theorem. Any simply connected open subset of $\mathbb{C}$ that is not all of $\mathbb{C}$ is biholomorphic to the unit disk. This is the deepest single statement of complex analysis: it says that all simply connected proper subsets of $\mathbb{C}$ are equivalent as complex-analytic objects.

Picard's theorems. Little Picard: an entire non-constant function takes every complex value, except possibly one. Great Picard: in any neighbourhood of an essential singularity, a holomorphic function takes every complex value, except possibly one, infinitely often.

Mittag-Leffler theorem. For any prescribed set of poles in $\mathbb{C}$ with assigned principal parts, there exists a meromorphic function on $\mathbb{C}$ realising exactly that pole-and-principal-part data.

Weierstrass factorisation theorem. Every entire function with prescribed zeros (and multiplicities) can be expressed as an infinite product of elementary factors.

Synthesis. This construction generalises the pattern fixed in 02.01.01 (topological space), with the symmetric data replaced by its skew or twisted analogue. Read in the opposite direction, the construction is dual to the metric story: complements and orthogonality are taken with respect to the bilinear datum of this unit, not a metric, and the resulting category of subobjects is the one the rest of the strand classifies. The central insight is that this datum identifies algebra with geometry: functions become vector fields, subspaces become quotients, and invariants become cohomology classes — and that identification is the engine driving every theorem downstream.

Full proof set [Master]

Detailed proofs of Goursat's theorem (the cornerstone of Cauchy theory), the residue theorem, Casorati-Weierstrass theorem, the Riemann mapping theorem (via normal families and the Montel theorem), and Picard's theorems are deferred to companion units. The key conceptual proofs of Cauchy's theorem and Liouville's theorem are given above.

Connections [Master]

• Topological space 02.01.01 — the underlying topology of $\mathbb{C}$ is the basis for all complex-analytic considerations.

• Continuous map 02.01.02 — holomorphic functions are continuous, but the converse is far from true.

• Riemann surface 06.03.01 — a 1-dimensional complex manifold, generalising domains in $\mathbb{C}$ to spaces where holomorphicity makes sense globally.

• Sheaf 04.01.01 — the sheaf of holomorphic functions ${\mathcal{O}}_X$ is the structure sheaf of any complex manifold.

• Riemann-Roch for compact Riemann surfaces 06.04.01 — the dimension formula for spaces of holomorphic sections of line bundles, the analytic version of algebraic Riemann-Roch.

• De Rham cohomology 03.04.06 — Dolbeault cohomology $H^{p,q}(X)=H^q(X;{\Omega }^p)$ refines de Rham cohomology on complex manifolds.

• Sheaf cohomology 04.03.01 — Dolbeault cohomology is sheaf cohomology of the sheaves ${\Omega }^p$ of holomorphic $p$-forms.

Historical & philosophical context [Master]

The systematic study of complex differentiable functions began with Augustin-Louis Cauchy in the early 1820s, who introduced contour integration and the integral formula. Bernhard Riemann's 1851 dissertation introduced the geometric perspective: holomorphic functions are conformal maps, and global complex analysis lives on Riemann surfaces.

The third foundational pillar came from Karl Weierstrass: holomorphic functions are power series. Weierstrass developed the theory of analytic continuation and entire functions through power-series methods, complementing Cauchy's integral methods and Riemann's geometric methods.

These three perspectives — Cauchy's analytic, Riemann's geometric, and Weierstrass's algebraic-power-series — are equivalent (modulo modest hypotheses), but each emphasises different aspects. The synthesis is what makes complex analysis simultaneously deep, elegant, and powerful.

The 20th century saw complex analysis generalise to several variables (the much harder several complex variables theory of Grauert-Remmert), to complex manifolds and Hodge theory (Kodaira, Hirzebruch), and to complex algebraic geometry through the GAGA principle (Serre, Grothendieck). The local concept "holomorphic function" remains foundational throughout.

Bibliography [Master]

• Ahlfors, Complex Analysis — the canonical introduction, masterful and elegant.
• Stein & Shakarchi, Complex Analysis (Princeton Lectures Vol. 2) — modern textbook with excellent applications.
• Conway, Functions of One Complex Variable I — comprehensive graduate-level treatment.
• Rudin, Real and Complex Analysis — rigorous, with strong measure-theoretic foundation.
• Forster, Lectures on Riemann Surfaces — extends complex analysis to its proper geometric setting.
• Hörmander, An Introduction to Complex Analysis in Several Variables — for the multivariate generalisation.