Analytic continuation and identity theorems
Use interior accumulation and overlapping analytic descriptions without confusing one Taylor disk with an entire function's domain.
On this page
The question for today
What you’ll understand
- State the identity theorem with connectedness and an interior accumulation point.
- Continue an analytic function using overlapping Taylor disks.
- Distinguish uniqueness of a continuation from its existence and path independence.
A Taylor series is one local description of a holomorphic function, not necessarily its largest possible domain. Overlapping disks can carry the same analytic information farther than the original radius allows. The identity theorem makes the overlap rigid: two holomorphic descriptions that agree on enough nearby points must agree throughout their connected common domain. This gives uniqueness where continuation is possible, while leaving separate questions about singularities, existence along a chosen path, and whether different paths return the same analytic branch.
Zeros cannot accumulate in the interior unless everything vanishes
Suppose a holomorphic function on a connected open domain is not identically zero. At any zero, its Taylor series has a first nonzero coefficient of some finite order. Factoring out the corresponding power leaves a holomorphic factor nonzero at that point and therefore nonzero in a smaller neighborhood. The original zero is isolated. If all Taylor coefficients vanished at one point, the function would vanish on a disk; connectedness then propagates that local vanishing through the domain.
Apply this argument to the difference of two functions. If they agree at distinct points accumulating at an interior point of their connected domain, they agree everywhere on that domain. Agreement on a real interval inside a complex domain is enough, even though the interval is not open in the plane. Agreement at finitely many points is not enough. An accumulation point on the boundary is also insufficient because the local Taylor argument needs a disk around the accumulation point contained in the domain.
Continue the function rather than summing outside a convergence disk
An analytic continuation extends a holomorphic description to a larger domain while agreeing on a nonempty open overlap. A chain of overlapping disks can implement this extension by repeatedly re-expanding at a new center. Each series is used only inside its own convergence disk. The original coefficients do not license substitution beyond their radius; instead, the compatible new series provides values there. A singularity can stop one route while another route remains available.
For a rational function, poles visibly constrain Taylor disks. A series centered at zero may stop at a pole on one side even though the function is perfectly regular far away on the other side. Re-expanding toward the regular side can reach those points. More generally, the identity theorem ensures that two putative extensions agree on a connected overlap when they match there, but it does not guarantee that an extension exists through a singular boundary point. Uniqueness is not a substitute for constructing or proving existence.
Boundary behavior and path dependence need their own analysis
Boundary accumulation can coexist with a nonzero holomorphic function. On the unit disk, has zeros at for integers n≥2, accumulating only at the excluded boundary point 1. There is no contradiction with isolated interior zeros. Likewise, a Taylor series failing to converge at a point does not by itself prove that the underlying function has a singularity there; the function might be represented by another continuation through a neighborhood of that point.
Continuation around holes can return a different local branch, as with logarithms. The identity theorem compares single-valued functions on a specified common domain, whereas a sequence of local continuations records a path. On suitable simply connected domains, a monodromy theorem can remove this path dependence when continuation exists along every path. Those existence and topology hypotheses will matter later. For now, keep track of the domain of each local description and justify agreement on every overlap.
Work it through
Reach a regular point beyond the first Taylor disk
Use a Taylor expansion centered at −1/2 to find the analytic continuation of at z=−1. Explain why directly substituting into the original series fails.
On its original unit disk the geometric series equals . At center a=−1/2, rewrite the denominator as .
The new geometric expansion is . It converges when , so it contains z=−1 and overlaps the original disk.
At z=−1 the new ratio is −1/3. Summing gives . The original series would become , whose ordinary partial sums do not converge.
The continued analytic value is 1/2, supplied by a compatible convergent series. This is not ordinary convergence of the original series at its boundary point.
The living notebook
Explore the model
Interactive lab · spectral
See why finitely many matching values are insufficient
Pause and Reset; set Mode count N=1 and Decay p=1, keeping time zero. Compare with a fresh Reset at N=2,p=1. Predict that both finite sine sums vanish at x=0,π,2π but differ between those locations. These are real traces of entire finite sine sums at the reset time. The graph illustrates finite agreement only; it does not test an infinite interior accumulation set or perform analytic continuation.
- Sum at x = π/2
- 0.867
- Smallest wavelength
- 2π / 5
- Highest-mode coefficient
- 0.2
A finite sum of sin(nx − t)/nᵖ. Playback changes t; increasing the cutoff includes finer spatial scales. This is not a proof about a continuum limit or renormalization. Parameter changes pause playback and recompute the model at the current elapsed time. Reset restores both parameters and returns time to zero. Playback starts paused. Adjust either parameter, then choose Play or Step.
Look forThe identity theorem needs an interior accumulation of distinct matching points, rather than a finite list of coincidences.
Pause & reason
Check your understanding
Your turn
Try it yourself
Functions f and g are holomorphic on the unit disk and satisfy for every integer n≥2. Must they agree everywhere? Would the same argument work with matching points instead?
Take a moment to make a prediction or work it out on paper. Open a hint when you need one.
Give me a hint
Locate the two accumulation points and check whether each belongs to the open domain.
Compare with the solution
The first points accumulate at 0, an interior point, so the identity theorem gives f=g throughout the disk. The second points accumulate at the boundary point 1, so that argument does not apply. Indeed f(z)=0 and agree at all yet are not identical; for example g(−1/2)=sin(2π/3) is nonzero.
Keep the idea
Take it with you
- Interior accumulation determines a holomorphic function on a connected domain.
- Overlapping Taylor disks can extend a function without extending one series' convergence radius.
- Uniqueness, existence, and path independence of continuation are distinct claims.
Read further
Original sources and deeper explanations to follow at your own pace.
Optional progress, saved in this browser when available. Revisit any lesson at any time.
Lesson not marked complete.