1 Identity theorem in complex analysis

1.1 Statement of the theorem

Let \(U\subset\mathbb{C}\) be a connected open set, and let \(f,g:U\to\mathbb{C}\) be holomorphic. If the set of points \(\{z\in U: f(z)=g(z)\}\) has an accumulation point in \(U\), then \(f\equiv g\) on \(U\). Equivalently, if a holomorphic function \(h\) has a zero set with an accumulation point in \(U\), then \(h\equiv 0\) on \(U\).

1.2 Role of holomorphicity

Holomorphicity (complex differentiability) imposes strong regularity: holomorphic functions are not merely differentiable but admit local power series expansions. This analytic structure is what makes “agreement on enough points” propagate to agreement everywhere in the connected domain.

1.3 Accumulation points and why they matter

An accumulation point \(z_0\in U\) means there are infinitely many distinct points of the agreement set approaching \(z_0\). Without such clustering, two holomorphic functions could coincide at isolated points yet still differ elsewhere; accumulation forces their difference to vanish to infinite order at a point, which is only possible if the difference is identically zero.

1.4 Connectedness of the domain

The theorem is stated for a connected domain because holomorphicity combined with identity constraints yields equality on the smallest connected region containing the accumulation point. On disconnected sets, one can have functions that agree on one component but not on another.

1.5 Examples illustrating the theorem

A typical example is the following: if \(f\) and \(g\) are holomorphic on a disc and \(f(z)=g(z)\) for infinitely many points having a limit inside the disc, then \(f=g\) throughout the disc. Another example uses zeros: a nonzero holomorphic function on a domain cannot have zeros accumulating in that domain. For instance, a function like \(h(z)=\sin z\) has infinitely many zeros, but they accumulate only at infinity; within any bounded region, the zeros are discrete.

2 Proof strategies

2.1 Proof via the zeros of a holomorphic function

A standard route reduces the problem to the case of a single function. Define \(h=f-g\). Then \(h\) is holomorphic, and the set where \(h=0\) has an accumulation point in \(U\). Showing \(h\equiv 0\) proves \(f\equiv g\).

2.1.1 Constructing the difference function

Because \(f\) and \(g\) are holomorphic on \(U\), their difference \(h=f-g\) is holomorphic on \(U\). Moreover, the equality \(f(z)=g(z)\) is equivalent to \(h(z)=0\). Thus the hypothesis translates directly into “the zeros of \(h\) have an accumulation point.”

2.1.1.1 Using power series expansions around an accumulation point

Let \(z_0\in U\) be an accumulation point of zeros of \(h\). By holomorphicity, \(h\) has a power series expansion near \(z_0\): \[ h(z)=\sum_{n=0}^{\infty} a_n (z-z_0)^n. \] If \(h\) is not identically zero, then there is a smallest index \(m\) such that \(a_m\neq 0\). Factoring yields \[ h(z)=(z-z_0)^m \cdot \phi(z), \] where \(\phi\) is holomorphic and \(\phi(z_0)\neq 0\). In a neighborhood of \(z_0\), \(\phi\) cannot vanish, so the only zero of \(h\) near \(z_0\) would be at \(z_0\) itself (with multiplicity \(m\)). This contradicts the existence of infinitely many zeros accumulating at \(z_0\). Hence all coefficients must be zero, so \(h\equiv 0\).

2.2 Alternative proof using analytic continuation ideas

Another approach uses analytic continuation in the form: a holomorphic function is determined by its values near any point in the connected domain. Starting from a neighborhood of the accumulation point, one shows that the Taylor series of \(h\) is forced to be zero, and then extends the equality along paths in the connected set. This emphasizes the rigidity of analytic functions as they propagate across the domain.

2.3 Using the maximum modulus principle variant

There are also proofs leveraging properties of \(h\). Consider \(h\) holomorphic and not identically zero. Then \(h\) is subharmonic, and on sufficiently small circles around a point one can derive constraints on how often \(h\) can vanish. If \(h\) has zeros arbitrarily close to an interior point, such constraints force \(h\) to be identically zero. While this route is less direct than the power-series argument, it aligns the theorem with broader maximum-principle phenomena in complex analysis.

3 Consequences and corollaries

3.1 Uniqueness of analytic continuation

If a holomorphic function is known on a set with an accumulation point and one extends it analytically to a larger connected domain, the extension is unique. In effect, two possible continuations must coincide on the overlap region, and the identity theorem upgrades “coincide on a non-discrete set” to “coincide everywhere on the connected overlap.”

3.2 Determination by boundary or interpolation data

Within a suitable analytic framework, specifying the function at an infinite set of points (subject to an accumulation condition) determines the entire function. For example, if two holomorphic functions on a connected domain agree on a sequence of interpolation nodes converging to a point in the domain, then their difference has zeros with an accumulation point and must vanish identically.

3.3 Consequences for zeros of holomorphic functions

A central corollary is structural: the zeros of a nontrivial holomorphic function are isolated inside the domain. More precisely, near any zero \(z_0\) of a nonzero holomorphic function, the function can be written as \((z-z_0)^m\phi(z)\) with \(\phi(z_0)\neq 0\). Thus the zero set cannot have accumulation points in the domain unless the function is identically zero.

3.4 Corollary forms with derivatives

Differentiated versions follow from applying the identity theorem to related holomorphic functions. For instance, if \(h\) has a zero of sufficiently high multiplicity at a point, then certain derivatives vanish there; conversely, if all derivatives at a point agree for two holomorphic functions and the point lies in the domain, then the two functions coincide. This reflects that the Taylor series around a point determines the function uniquely.

4.1 Identity theorem for real-analytic functions

A parallel statement holds for real-analytic functions: if two real-analytic functions agree on a set that has an accumulation point within an interval (or, more generally, within a connected real-analytic setting), then they agree everywhere on the connected region. The proof uses the analogous power-series expansion (now in real variables) and the rigidity of convergent Taylor series.

4.2 Identity theorem in several complex variables

In \(\mathbb{C}^n\), the identity theorem remains valid for holomorphic functions on connected domains. If two holomorphic functions agree on a set having an appropriate type of accumulation (often described using that the set has a nontrivial limit structure), then they must coincide on the entire connected domain. The geometry of zero sets is subtler in higher dimensions, but the core principle—analyticity prevents “local agreement without global equality”—persists.

4.3 Versions for holomorphic functions on domains with multiple components

If \(U\) is not connected, the conclusion becomes componentwise: equality holds on each connected component that contains the accumulation point. This clarifies that the connectedness hypothesis is not technical excess; it directly matches the possibility of independent holomorphic behavior across different components.

4.4 Comparison with the zero set structure of analytic functions

Analytic functions in the complex plane exhibit a stark restriction: zeros are isolated unless the function is identically zero. This contrasts with smooth (infinitely differentiable) functions, where a nontrivial function may have a zero set with accumulation points without being forced to vanish identically. The identity theorem can therefore be read as a manifestation of the special “analytic zero rigidity” of holomorphic functions.

5 Applications

5.1 Analytic continuation of germs and global extension

In complex analysis, one often starts with a holomorphic “germ” at a point, meaning the function is known only in some neighborhood. When that local information extends along a path, the identity theorem ensures the extension is compatible: if two extended versions overlap, they agree on the overlap and hence must match globally on the connected region of the continuation.

5.2 Determining parameters in analytic families

Suppose \(f(z,\lambda)\) is holomorphic in \(z\) for each parameter \(\lambda\), and the dependence on \(\lambda\) is such that differences remain holomorphic in a relevant sense. If two parameter choices yield the same function values at a set of \(z\)-points with accumulation in the domain, the identity theorem can force equality of the corresponding functions. This is commonly used to show that certain analytic parameters are uniquely determined by observed data.

5.3 Rigidity arguments in conformal mapping contexts

Conformal maps are holomorphic and have additional structure (e.g., nonvanishing derivatives under normalization). When two conformal maps coincide on a set with accumulation in the domain, the identity theorem implies they are the same map. This supports uniqueness results for mappings determined by boundary behavior or normalization conditions.

5.4 Use in proving equality of special functions defined analytically

Many special functions are defined via analytic formulas, differential equations, or integral representations that produce holomorphic functions on particular domains. If two candidate definitions yield holomorphic functions agreeing on a set with an accumulation point, the identity theorem allows one to conclude they are identical as functions, giving a rigorous bridge between alternative representations.

6 Common pitfalls and limitations

6.1 Failure for non-analytic (smooth but not analytic) functions

The theorem is specific to holomorphic (or real-analytic) functions. Smooth functions can vanish on complicated sets with accumulation points without being identically zero. Consequently, replacing holomorphicity by mere differentiability destroys the rigidity that drives the identity theorem.

6.2 What happens without an accumulation point

If the set where \(f=g\) is discrete with no accumulation in the domain, the conclusion can fail. Two holomorphic functions can agree at isolated points yet differ elsewhere; the identity theorem requires the stronger “local clustering” property.

6.3 Dependence on domain assumptions (open/connected)

The domain must be open for holomorphicity to imply power-series expansions, and connectedness is needed for global equality on the whole region. If the function is defined on a non-open set, or the domain is split into components, the propagation of equality may stop at component boundaries or fail altogether due to lack of analytic control.