Analytic Continuation

Motivation: A function "knows" itself everywhere

Holomorphic functions possess a remarkable property: knowing the function in a small domain allows one to extend it to everything permitted by analyticity. This is called analytic continuation—and it is unique. It is precisely this principle that allows the “definition” of the Riemann ζ-function beyond the half-plane Re s > 1, where it is given by the original series.

Uniqueness Principle

Theorem: If f and g are holomorphic in a connected domain D and coincide on a set with a limit point in D, then f ≡ g in D.

Corollary (analytic continuation principle): If f is holomorphic in D₁, g is holomorphic in D₂, D₁ ∩ D₂ ≠ ∅ is connected, and f = g on D₁ ∩ D₂, then g is the unique analytic continuation of f to D₂.

Physical meaning: There do not exist two different holomorphic functions that coincide even on a tiny segment. “A holomorphic object is an indivisible whole.”

Continuation Along a Path

A function f, defined in the circle K₀, is continued along a path γ via a chain of overlapping circles K₀, K₁, ..., Kₙ, where in each Kᵢ a "branch" fᵢ is defined, coinciding with fᵢ₋₁ on the intersection.

Monodromy: Continuation along two paths with the same endpoints may yield different results if the paths “enclose” a singular point.

Riemann Surfaces and Multivaluedness

Example — √z: Continue around a circle around 0. Starting with √r·e^{iθ/2} at θ = 0 and traversing 2π, we arrive at √r·e^{iπ} = −√r—another "branch." The Riemann surface for √z is a two-sheeted covering of the sphere: two sheets, glued along the cut [0,+∞).

Logarithm: The Riemann surface for ln z is an infinitely sheeted covering: after each circuit around 0 we move to a new sheet, increasing Im(ln z) by 2π.

The Riemann Zeta Function

Definition for Re s > 1: ζ(s) = Σₙ₌₁^∞ 1/nˢ (absolutely convergent).

Analytic continuation: ζ(s) is continued to ℂ \ {1} with a single pole (simple) at s = 1. This is nontrivial: the original series diverges for Re s ≤ 1.

Functional equation: ζ(s) = 2^s π^{s–1} sin(πs/2) Γ(1–s) ζ(1–s).

Riemann Hypothesis: All nontrivial zeros of ζ(s) (those not at s = –2, –4, ...) lie on the “critical line” Re s = 1/2. This is one of the seven Millennium Prize Problems, for which a reward of $1 million is offered.

Numerical Example

Problem: Continue the function f(z) = Σₙ₌₀^∞ zⁿ from the disk |z| < 1 to the point z = 2 via an intermediate disk.

Step 1. For |z| < 1: f(z) = 1/(1−z) (the sum of the geometric series).

Step 2. Analytic continuation is simply the function 1/(1−z), defined on ℂ \ {1}. At the point z = 2: f(2) = 1/(1−2) = –1.

Step 3. This seems paradoxical: the series Σzⁿ diverges at |z| = 2, but the analytic continuation yields –1. However, it is precisely in this sense that “formally” 1 + 2 + 4 + 8 + … = –1 (an analogous “shift” appears in ζ-functional computations).

Step 4. Through the intermediate center z₀ = 0.5: the Taylor series of 1/(1−z) at the point 0.5 is Σₙ₌₀^∞ (z–0.5)ⁿ/(0.5)^{n+1}, convergent in |z–0.5| < 0.5. This covers z = 0.9. The next disk—centered at 0.9—covers up to 1.4, and so on. Continuing the chain, we reach z = 2 (bypassing z = 1).

Real-World Application

Analytic continuation is used in quantum field theory for “renormalization”: divergent integrals are regularized in the complex domain, and the physical result is obtained by analytic continuation to real parameters.

Additional Aspects

Analytic continuation is implemented numerically through the Padé approximation method: a power series is replaced by a rational fraction of the same order, which confidently “circumvents” poles and works outside the radius of convergence of the original series. In quantum field theory, analytic continuation via the complex momentum plane (Wick rotation t → iτ) turns oscillatory integrals into convergent Gaussian integrals—without this technique, perturbation theory calculations are impossible. In number theory, the Riemann zeta-function ζ(s) = Σ 1/n^s, defined only for Re s > 1, is analytically continued to the entire plane with a single pole at s = 1; it is precisely this continuation that enables discussion of “zeros of ζ” and formulation of the Riemann Hypothesis.

Connection with Other Areas of Mathematics

In the theory of differential equations, analytic continuation is embedded in the very construction of solutions. The classical example—analytic solutions of linear equations with analytic coefficients: the Cauchy–Kovalevskaya theorem guarantees a local solution, and a global description, including monodromy, is obtained by continuation along paths. In constructing fundamental solutions, one transitions from a local power series to a global holomorphic (or multivalued) function on a Riemann surface.

In algebraic geometry, the concept of analytic continuation arises via the GAGA principle: on a compact complex algebraic variety holomorphic functions coincide with regular ones. This allows translation of problems about function continuation into the language of ideals and morphisms of schemes. Riemann’s theorem on the extendability of meromorphic functions on a compact Riemann surface (every holomorphic function on such a surface is constant, and every meromorphic function has only finitely many poles) is the analytic analogue of algebraic facts about rational functions on curves.

Topological aspects manifest through monodromy and coverings: continuation along different paths generates a representation of the fundamental group of the domain in the automorphism group of the solution space (the monodromy group). In the theory of fiber bundles this is described as flat connection and parallel transport. The work of Weyl and Cartan on differential forms and connections builds this bridge between complex analysis and the topology of smooth manifolds.

In probability problems, analytic continuation appears in connection with characteristic functions and generators of Markov processes: for example, continuation of heat kernel densities in complex time is related to the method of images and functional integrals in physics. In numerical methods, apart from Padé approximants, stabilizing schemes are used for solving inverse problems, where analytic continuation essentially represents an ill-conditioned operation of reconstructing a function from its values on a limited domain.

Historical Background and Development of the Idea

The birth of the idea of analytic continuation is associated with the works of Cauchy (1820s–1830s), where the principle of uniqueness of holomorphic functions is already present. Riemann, in his 1851 dissertation on representing functions of a complex variable via harmonic potentials, introduces surfaces on which multivalued functions become single-valued—the prototype of modern Riemann surfaces. At the end of the 19th century, Weierstrass formalizes analytic functions as power series and considers their continuation via chains of convergence disks. Poincaré and Cauchy actively use continuation along paths in studying linear differential equations and introduce the concept of monodromy for analytic solutions. The decisive step is linked to the works of Erdmann and Schwarz on the principle of identity and uniqueness theorems, which led to the modern formulation of analytic continuation via coincidence on a set with a limit point.