Alphabeta Math
Session-authored (Fable 5 assisted)
How statement and proof provenance work

The first chip identifies the source of the statement or construction; the second identifies the source of its local proof or verification.

  • Literature-sourced: the exact statement appears in a cited source; only wording and notation differ.
  • AI-adapted: a semantically identical restatement of literature-sourced material, modulo indexing, notation, and boundary cases adopted by the library.
  • AI-generated: a genuinely novel statement formulated by AI, with no source for the claim itself.

These labels describe origin, not correctness: citations and verification chips remain separate evidence.

11 results · all verified · 0 also independently AI-judged
Every result on this page is machine-checked by a proof checker and read in full and owner-audited; the judge is an additional, independent cross-model AI review of the proofs. The 11 not AI-judged were verified by owner audit (typically over a confirmed judge false positive), not failures.

The Argument Principle and Rouché's Theorem

1 · Prerequisites

2 · Summary

This page turns the residue theorem into zero and pole counting. The first half defines the logarithmic derivative and the winding-weighted zero and pole sums, then proves the argument principle in both its residue-counting form and its geometric image-winding form. That geometric reading is the bridge to Rouché's theorem: the boundary homotopy keeps the image winding number fixed, so the interior zero count cannot change.

The second half packages the standard consequences that the rest of the complex analysis track uses later: counting preimages of a target value, weighting the count by a holomorphic test function, the local stability of zero multiplicity under perturbation, the Hurwitz zero-free and injective-limit theorems, the agreement with the earlier open-mapping and local-degree pages, and the contour formula that recovers a locally single-valued inverse branch.

3 · Logical flowchart

4 · Definitions, theorems and proofs

DefinitionDefinition: Literature-sourcedProof: Not applicablejudge pass (gpt-5.6-terra)audited 2026-08-28Open item page →

The logarithmic derivative of a meromorphic function

Definition

Let ΩC be a complex domain, let f be meromorphic on Ω, let Z(f) be the zero set of f, and let P(f) be its pole set. On the open set

Ω(Z(f)P(f))

the function f is holomorphic and nonzero, so the quotient

ff

is holomorphic there. This quotient is the logarithmic derivative of f.

Remarks

The logarithmic derivative is not defined at a zero or a pole of f by the displayed quotient itself. At every zero of finite positive order and at every pole, the next lemma identifies its principal part and shows that the resulting singularity is simple. The identically zero function is excluded from that local conclusion because its quotient f/f is defined nowhere.

DefinitionDefinition: Literature-sourcedProof: Not applicableaudited 2026-08-28Open item page →

Zero and pole counts weighted by multiplicity and winding number

Definition

Let ΩC be open, let f be meromorphic on Ω, and let Γ be admissible for the residue theorem in Ω. Assume also that f is not identically zero on any connected component of Ω and has no zeros on Γ, so every zero has finite positive order and every index n(Γ,a) of Integration over a complex chain and the index of a chain is defined at every zero or pole of f.

Here, and in the argument-principle results that use this definition, meromorphic on an open set means meromorphic on every connected component in the sense of Meromorphic functions on a plane domain. The zero and pole sets are the unions of the corresponding componentwise sets.

Whenever only finitely many zeros and poles of f have nonzero index with respect to Γ, define the weighted zero count

Z(f,Γ):=aZ(f)n(Γ,a)orda(f)

and the weighted pole count

P(f,Γ):=bP(f)n(Γ,b)ordbpole(f),

where orda(f) is the zero order from The order of a zero of a holomorphic function and ordbpole(f) is the positive pole order at b.

Remarks

These are finite sums only after a finiteness argument. Under the argument principle hypotheses, that finiteness comes from applying the residue theorem to the logarithmic derivative.

LemmaStatement: Literature-sourcedProof: AI-adaptedprecheck passaudited 2026-08-28Open item page →

The logarithmic derivative has residue equal to local order

Statement

Let f be meromorphic on a neighbourhood of aC.

  1. If a is a zero of f of order m1, then Res ⁣(ff,a)=m.
  2. If a is a pole of f of order m1, then Res ⁣(ff,a)=m.

In either case f/f has a simple pole at a.

Facts & Assumptions

Given: A meromorphic function f on a neighbourhood of a.

[L1]

A holomorphic function has a zero of order m at a exactly when it factors locally as (za)mh(z) with h holomorphic and h(a)0 (The order of a zero is the exponent in its local holomorphic factorization).

[L2]

A pole of order m is exactly a point where 1/f extends holomorphically across a and has a zero of order m there (Characterizations of poles).

[L3]

Holomorphic quotients and products obey the usual derivative rules, and a holomorphic function is continuous (Linearity, product, reciprocal, and quotient rules for complex derivatives, Complex differentiability at a point implies continuity there).

Proof

technique · direct
1.1

Suppose first that a is a zero of f of order m. By [L1], on a disc about a one has f(z)=(za)mh(z) with h holomorphic and h(a)0.

givenL1
1.2

Suppose instead that a is a pole of f of order m. By [L2], 1/f=(za)mh locally for some holomorphic h with h(a)0. Shrinking as before, h is nowhere zero.

givenL2L3
2.1

By continuity in [L3], shrink the disc so that h is nowhere zero there. Differentiating the factorization from step 1.1 and dividing by (za)mh(z) gives f(z)f(z)=mza+h(z)h(z). The second term is holomorphic by [L3], so the residue is m and the pole is simple.

step 1.1L3algebra
3.1

From step 1.2 one has f(z)=(za)mh(z)1. Differentiating and dividing by f yields f(z)f(z)=mzah(z)h(z). Again the second term is holomorphic by [L3], so the residue is m and the pole is simple.

step 1.2L3algebra
TheoremStatement: Literature-sourcedProof: AI-adaptedprecheck passaudited 2026-08-28Open item page →

The argument principle for an admissible null-homologous cycle

Statement

Let ΩC be open, let f be meromorphic on Ω, and let Γ be admissible for the residue theorem in Ω. Suppose in addition that f is not identically zero on any connected component of Ω and that f(z)0 for every zΓ. Then

12πiΓf(z)f(z)dz=Z(f,Γ)P(f,Γ),

where the weighted zero and pole counts are those of Zero and pole counts weighted by multiplicity and winding number.

Only finitely many terms in those weighted counts are nonzero.

As in Zero and pole counts weighted by multiplicity and winding number, meromorphicity on the possibly disconnected open set Ω is understood componentwise.

Facts & Assumptions

Given: An open set Ω, a meromorphic function f on Ω that is not identically zero on any connected component, and an admissible cycle Γ in Ω such that f has no zero on Γ.

[L1]

Away from the zeros and poles of f, the logarithmic derivative f/f is holomorphic (The logarithmic derivative of a meromorphic function).

[L2]

At a zero of order m, the logarithmic derivative has residue m, and at a pole of order m it has residue m (The logarithmic derivative has residue equal to local order).

[L3]

A meromorphic function admissible for a cycle has only finitely many poles with nonzero index (Only finitely many singularities contribute to the residue sum of an admissible cycle).

[L4]

The residue theorem for an admissible null-homologous cycle reads Γg(z)dz=2πicn(Γ,c)Res(g,c) with only finitely many nonzero terms (The residue theorem for a null-homologous cycle).

Proof

technique · direct
1.1

By [L1], the function g:=f/f is holomorphic away from the zeros and poles of f. By [L2], every zero or pole of f becomes a simple pole of g. Because f has neither zeros nor poles on Γ, the cycle Γ is admissible for g as well.

givenL1L2
2.1

Applying [L3] to g shows that only finitely many zeros or poles of f have nonzero index with respect to Γ. Therefore the sums defining Z(f,Γ) and P(f,Γ) are finite.

step 1.1L3
3.1

By [L4] applied to g, 12πiΓf(z)f(z)dz=cn(Γ,c)Res ⁣(ff,c), where c ranges over the poles of g. Splitting those poles into zeros and poles of f and then using [L2] turns the right-hand side into Z(f,Γ)P(f,Γ).

step 1.1step 2.1L2L4
TheoremStatement: Literature-sourcedProof: AI-adaptedprecheck passaudited 2026-08-28Open item page →

The argument-principle integral is the winding number of the image cycle

Statement

Let γ:[a,b]C be a closed complex contour, let f be meromorphic on a neighbourhood of γ, and suppose f(z)0 for every zγ. Then fγ is a closed complex contour with 0(fγ), and

12πiγf(z)f(z)dz=n(fγ,0).

Equivalently, if θ is any continuous argument of fγ, then

12πiγf(z)f(z)dz=θ(b)θ(a)2π.

If γ is also admissible and null-homologous in a larger open set on which f is meromorphic, then the same integer equals Z(f,γ)P(f,γ) by The argument principle for an admissible null-homologous cycle.

Facts & Assumptions

Given: A closed complex contour γ, a meromorphic function f on a neighbourhood of γ, and f(z)0 on γ.

[L1]

The winding number of a closed contour about a point off its trace is n(η,p)=12πiηdwwp (The winding number of a closed contour about a point off its trace).

[L2]

The winding number is also the normalized increment of any continuous argument (The winding number is the increment of a continuous argument divided by 2π).

[L3]

A contour missing the origin admits a continuous logarithm, unique up to a constant in 2πiZ (Every contour missing a point admits a continuous logarithm, unique up to a constant in 2πiZ).

[L4]

A holomorphic nonvanishing function on a disc has a holomorphic logarithm, and that logarithm has derivative f/f (A nonvanishing holomorphic function on a disc has a holomorphic logarithm, A holomorphic logarithm is a primitive of the logarithmic derivative).

Proof

technique · direct
1.1

Since f is continuous on the compact set γ and never vanishes there, fγ is a closed complex contour whose trace misses 0. By [L3], choose a continuous logarithm λ of fγ. Cover γ by finitely many open discs U1,,UN on which f has no zeros, and then subdivide γ into consecutive subcontours γj whose traces lie in those discs.

givenL3choose
2.1

Fix j. On Uj, [L4] gives a holomorphic logarithm Lj of f, with Lj=f/f. Along the trace of γj, both Ljγj and λγj are continuous logarithms of fγj, so [L3] makes their difference constant. Therefore λ(tj)λ(tj1)=Lj(γ(tj))Lj(γ(tj1))=γjf(z)f(z)dz, where the last equality is [L5] applied to the primitive Lj.

L3L4L5step 1.1
3.1

Summing the equalities of step 2.1 over the subdivision and using the additivity from [L5] gives γf(z)f(z)dz=λ(b)λ(a). Now [L1] and [L2] applied to the contour fγ identify the same increment with both 2πin(fγ,0) and i(θ(b)θ(a)), so the two displayed formulas follow.

L1L2L5step 2.1algebra
CorollaryStatement: Literature-sourcedProof: AI-adaptedprecheck passaudited 2026-08-28Open item page →

The argument principle counts preimages of a target value

Statement

Let ΩC be open, let f be meromorphic on Ω, let Γ be admissible for the residue theorem in Ω, and let wC satisfy f(z)w for every zΓ. Then

12πiΓf(z)f(z)wdz=Nw(f,Γ)P(f,Γ),

where

Nw(f,Γ):=f(a)=wn(Γ,a)orda(fw)

is the weighted multiplicity count of the preimages of w, and P(f,Γ) is the weighted pole count of f.

In particular, if f is holomorphic on Ω, then the pole term vanishes and the integral counts the preimages of w with multiplicity.

Facts & Assumptions

Given: A meromorphic function f on an open set Ω, an admissible cycle Γ, and a complex number w with f(z)w on Γ.

[L1]

The argument principle applied to a meromorphic function g gives 12πiΓg(z)g(z)dz=Z(g,Γ)P(g,Γ) (The argument principle for an admissible null-homologous cycle).

[L2]

Derivatives ignore constants, so (fw)=f (Linearity, product, reciprocal, and quotient rules for complex derivatives).

Proof

technique · direct
1.1

Put g:=fw. Then g is meromorphic on Ω, has the same poles as f, and has no zero on Γ by the hypothesis on w. Its zeros are exactly the points a with f(a)=w.

givenL2
2.1

Applying [L1] to g and then using [L2] gives 12πiΓf(z)f(z)wdz=12πiΓg(z)g(z)dz=Z(g,Γ)P(g,Γ).

step 1.1L1L2
3.1

By step 1.1, the zero count Z(g,Γ) is exactly Nw(f,Γ) and the pole count P(g,Γ) is exactly P(f,Γ). Substituting that into step 2.1 proves the formula. If f is holomorphic, then it has no poles, so P(f,Γ)=0.

step 1.1step 2.1
TheoremStatement: Literature-sourcedProof: AI-adaptedprecheck passaudited 2026-08-28Open item page →

The weighted argument principle

Statement

Let ΩC be open, let f be meromorphic on Ω, let Γ be admissible for the residue theorem in Ω, suppose f(z)0 for every zΓ, and let g be holomorphic on Ω. Then

12πiΓg(z)f(z)f(z)dz=aZ(f)n(Γ,a)orda(f)g(a)bP(f)n(Γ,b)ordbpole(f)g(b),

and only finitely many terms are nonzero.

Facts & Assumptions

Given: An open set Ω, a meromorphic function f on Ω, an admissible cycle Γ with f0 on Γ, and a holomorphic function g on Ω.

[L1]

The logarithmic derivative has residue m at a zero of order m and residue m at a pole of order m (The logarithmic derivative has residue equal to local order).

[L2]

The unweighted argument principle already shows that only finitely many zeros and poles of f have nonzero index with respect to Γ (The argument principle for an admissible null-homologous cycle).

[L3]

The residue theorem sums the indexed residues of an admissible meromorphic function over Γ (The residue theorem for a null-homologous cycle).

Proof

technique · direct
1.1

Put h(z):=g(z)f(z)/f(z). Away from the zeros and poles of f, the function f/f is holomorphic, so h is holomorphic there as well. At a zero or pole c of f, the function g is holomorphic and therefore admits the expansion g(z)=g(c)+(zc)u(z) near c for some holomorphic u. Multiplying that by the principal-part decomposition from [L1] shows Res(h,c)=g(c)Res ⁣(ff,c).

givenL1algebra
2.1

Step 1.1 and [L1] therefore give Res(h,a)=g(a)orda(f) at each zero a of f, and Res(h,b)=g(b)ordbpole(f) at each pole b of f. By [L2], only finitely many such points have nonzero index with respect to Γ.

step 1.1L1L2
3.1

Applying [L3] to h and substituting the residue values from step 2.1 gives the displayed weighted sum formula.

step 2.1L3
TheoremStatement: Literature-sourcedProof: AI-adaptedprecheck passaudited 2026-08-28Open item page →

Rouche's theorem in the classical strict-inequality form

Statement

Let ΩC be open, let γ be a closed complex contour that is null-homologous in Ω, and let f,g be holomorphic on Ω. If

f(z)g(z)<g(z)(zγ),

then f and g have the same weighted number of zeros with respect to γ.

In particular, if γ is the positively oriented boundary of a Jordan domain, then f and g have the same number of zeros inside γ, counted with multiplicity.

Facts & Assumptions

Given: An open set Ω, a closed complex contour γ that is null-homologous in Ω, and holomorphic functions f,g on Ω satisfying fg<g on γ.

[L1]

For a closed contour on which a meromorphic function does not vanish, the integral of f/f is the winding number of the image contour about 0 (The argument-principle integral is the winding number of the image cycle).

[L2]

The argument principle at w=0 counts zeros of a holomorphic function with multiplicity and no pole term (The argument principle counts preimages of a target value).

[L3]

If φ(ζ,t) is continuous in (ζ,t) and holomorphic in the complex parameter t, then γφ(ζ,t)dζ is holomorphic in t (A contour integral of a jointly continuous, parameter-holomorphic integrand is holomorphic).

[L4]

Proof

technique · direct
1.1

Because γ is compact and fg<g there, the ratio fg/g has a maximum q<1 on γ. Choose ε>0 with (1+ε)q<1. Then for every complex t with t<1+ε and every zγ, g(z)+t(f(z)g(z))g(z)tf(z)g(z)<(1+ε)qg(z)<g(z), so ht(z):=g(z)+t(f(z)g(z)) never vanishes on γ.

givenchoosealgebra
2.1

For fixed z, the function φz(t):=ht(z)ht(z)=g(z)+t(f(z)g(z))g(z)+t(f(z)g(z)) is holomorphic on the disc t<1+ε by step 1.1. Therefore J(t):=12πiγht(z)ht(z)dz is holomorphic there by [L3]. For real t[0,1], step 1.1 and [L1] give J(t)=n(htγ,0), and [L4] makes that an integer. Hence J is an integer-valued holomorphic function on a connected open disc, so it is constant.

step 1.1L1L3L4
3.1

Since h0=g and h1=f, step 2.1 gives 12πiγg(z)g(z)dz=12πiγf(z)f(z)dz. Applying [L2] to both sides shows that f and g have the same weighted zero count with respect to γ.

step 2.1L2
RemarkRemark: Literature-sourcedProof: Not applicablejudge pass (gpt-5.6-terra)audited 2026-08-28Open item page →

Rouche gives the standard leading-term proof of the fundamental theorem of algebra

Remark

Let

p(z)=anzn+an1zn1++a0,an0.

On the circle z=R one has

an1zn1++a0an1Rn1++a0,

so for sufficiently large R the lower-degree tail is strictly smaller than anRn=anzn. Rouché's theorem therefore gives the same number of zeros for p and its leading term anzn inside z<R, namely n counting multiplicity.

This is exactly the standard leading-term proof that every nonconstant complex polynomial has roots and, more sharply, that a degree-n polynomial has exactly n roots counted with multiplicity, agreeing with the canonical result A complex polynomial of degree n has exactly n roots counted with multiplicity.

CorollaryStatement: Literature-sourcedProof: AI-adaptedprecheck passaudited 2026-08-28Open item page →

Small perturbations preserve the total local zero multiplicity

Statement

Let f be holomorphic on a neighbourhood of the closed disc D(a,r), and suppose f has no zero on the circle za=r. Let m be the total multiplicity of the zeros of f in D(a,r). If g is holomorphic on a neighbourhood of D(a,r) and

g(z)f(z)<f(z)(za=r),

then g has exactly m zeros in D(a,r), counted with multiplicity.

In particular, if a is an isolated zero of f of order m and the disc is chosen so that a is the only zero of f in D(a,r), then every such perturbation g has exactly m zeros in D(a,r) counted with multiplicity.

Facts & Assumptions

Given: A holomorphic function f on a neighbourhood of D(a,r), a holomorphic function g on the same neighbourhood, and gf<f on za=r.

[L1]

Rouché's theorem gives equal zero counts inside a closed contour when the strict boundary inequality holds (Rouche's theorem in the classical strict-inequality form).

Proof

technique · direct
1.1

Let γ(t)=a+reit for 0t2π. The hypothesis says gf<f on γ, and f has no zero there.

given
2.1

Applying [L1] to the contour γ shows that f and g have the same weighted zero count inside za<r. Because both are holomorphic, there is no pole term, so that weighted count is exactly the total multiplicity of the interior zeros. Hence g has as many zeros in the disc, counted with multiplicity, as f does.

step 1.1L1
3.1

The isolated-zero specialization is the case where that total multiplicity for f is the single local order m at a.

step 2.1
TheoremStatement: Literature-sourcedProof: AI-adaptedprecheck passaudited 2026-08-28Open item page →

Locally uniform convergence preserves the total multiplicity near an isolated zero

Statement

Let ΩC be open, let fn:ΩC be holomorphic, and suppose fnf locally uniformly on Ω. Let aΩ be an isolated zero of f of multiplicity m1. Then there is r>0 such that D(a,r)Ω, the function f has no zero on za=r, and for all sufficiently large n the function fn has exactly m zeros in D(a,r) counted with multiplicity.

Facts & Assumptions

Given: Holomorphic functions fn on an open set Ω converging locally uniformly to f, and an isolated zero a of f of multiplicity m.

[L1]

A strict boundary perturbation preserves the total zero multiplicity in the disc (Small perturbations preserve the total local zero multiplicity).

Proof

technique · direct
1.1

By the isolated-zero hypothesis, choose r>0 with D(a,r)Ω such that a is the only zero of f in D(a,r). Then f is a positive continuous function on the circle za=r, so it has a positive minimum there. Call that minimum η.

given
2.1

Because fnf locally uniformly and the circle za=r is compact, there is N such that fn(z)f(z)<ηf(z)(za=r, nN). Applying [L1] on that circle shows that for every nN, the function fn has the same total zero multiplicity in D(a,r) as f, namely m.

step 1.1L1
3.1

Step 2.1 is exactly the claimed persistence of the local multiplicity.

step 2.1
TheoremStatement: Literature-sourcedProof: AI-adaptedprecheck passaudited 2026-08-28Open item page →

Hurwitz's zero-free limit theorem

Statement

Let ΩC be a complex domain, let each fn:ΩC be holomorphic and nowhere zero, and suppose fnf locally uniformly on Ω. Then either f0 on Ω, or f is nowhere zero on Ω.

Facts & Assumptions

Given: A complex domain Ω, holomorphic nowhere-zero functions fn on Ω, and locally uniform convergence fnf.

[L2]

Near an isolated zero of the limit, sufficiently late approximants have the same total zero multiplicity (Locally uniform convergence preserves the total multiplicity near an isolated zero).

Proof

technique · direct
1.1

By [L1], the limit function f is holomorphic on Ω. If f0, the first alternative of the statement holds.

givenL1
2.1

Assume f≢0 and suppose toward a contradiction that f(a)=0 at some point aΩ. Then a is an isolated zero of f, so [L2] gives a disc about a on which every sufficiently large fn has at least one zero. That contradicts the hypothesis that every fn is nowhere zero.

step 1.1L2assume-contradischarge-contradiction
3.1

Therefore, in the nonzero branch of step 1.1, the function f has no zeros on Ω. This is the second alternative.

step 1.1step 2.1
TheoremStatement: Literature-sourcedProof: AI-adaptedprecheck passaudited 2026-08-28Open item page →

A locally uniform limit of injective holomorphic functions is injective or constant

Statement

Let ΩC be a complex domain, let each fn:ΩC be holomorphic and injective, and suppose fnf locally uniformly on Ω. Then f is injective or constant.

Facts & Assumptions

Given: A complex domain Ω, injective holomorphic functions fn on Ω, and locally uniform convergence fnf.

[L2]

A nonzero holomorphic function on a complex domain has isolated zeros (Zeros of a nonzero holomorphic function are isolated).

[L3]

Near an isolated zero of the limit, sufficiently late approximants preserve the total zero multiplicity (Locally uniform convergence preserves the total multiplicity near an isolated zero).

Proof

technique · contradiction
1.1

By [L1], the limit f is holomorphic. Assume it is not constant.

givenL1assume-contra
2.1

Suppose toward a contradiction that f is not injective. Then there are distinct points a,bΩ with f(a)=f(b)=:w. Because f is nonconstant, the function fw is not identically zero, so [L2] makes both a and b isolated zeros of fw. Choose disjoint closed discs D(a,ra),D(b,rb)Ω containing no other zeros of fw.

step 1.1L2choose
3.1

Apply [L3] to the sequence fnw on each of those discs. For all sufficiently large n, the function fnw has at least one zero in D(a,ra) and at least one zero in D(b,rb). Since the discs are disjoint, those are two distinct preimages of w, contradicting injectivity of fn.

step 2.1L3discharge-contradiction
4.1

The contradiction in step 3.1 shows that a nonconstant limit must be injective. Therefore every limit is injective or constant.

step 1.1step 3.1discharge-contradiction
RemarkRemark: Literature-sourcedProof: Not applicableaudited 2026-08-28Open item page →

The argument principle recovers the open mapping theorem

Remark

If f is nonconstant and holomorphic on a complex domain and aΩ, then ff(a) has a zero of positive multiplicity at a. Choose a small circle around a on which ff(a) does not vanish. The preimage-count corollary says that for every w sufficiently close to f(a) the function fw has the same positive number of zeros inside that circle. So every value near f(a) is attained nearby, and the image is open.

This reproduces the already-published open mapping theorem Open mapping theorem for holomorphic functions. The earlier page keeps its local-normal-form proof because the reading order needs that theorem before the argument principle exists.

RemarkRemark: Literature-sourcedProof: Not applicableaudited 2026-08-28Open item page →

Argument-principle multiplicity agrees with the earlier local degree

Remark

Fix a nonconstant holomorphic map f and a point a of local degree m=degaf. The earlier local-sheet theorem A local degree-m holomorphic map has m nearby sheets says that every nearby value other than f(a) has exactly m nearby preimages. On the other hand, the argument-principle preimage count on a small circle around a counts the zeros of fw in that same disc, with multiplicity, and therefore gives the same integer m.

So the multiplicity seen analytically by the argument principle agrees with the local degree already built from the normal form. This is an agreement remark, not a replacement of the earlier construction.

TheoremStatement: Literature-sourcedProof: AI-adaptedprecheck passaudited 2026-08-28Open item page →

A contour formula for a locally single-valued holomorphic inverse

Statement

Let ΩC be open, let f be holomorphic on Ω, let Γ be a closed complex contour null-homologous in Ω, and let wC satisfy f(z)w for every zΓ. Suppose f(z)=w has exactly one solution a inside Γ, that solution is simple, and n(Γ,a)=1. Then

a=12πiΓζf(ζ)f(ζ)wdζ.

Thus, on a contour enclosing exactly one simple preimage branch, the inverse value is recovered by a contour integral.

Facts & Assumptions

Given: A holomorphic function f on an open set Ω, a closed null-homologous contour Γ, and a value w satisfying the hypotheses of the statement.

[L1]

The weighted argument principle multiplies each zero contribution by the value of the holomorphic test function there (The weighted argument principle).

[L2]

The preimage-count corollary identifies the zeros of fw inside Γ (The argument principle counts preimages of a target value).

Proof

technique · direct
1.1

Because f is holomorphic, the meromorphic function fw has no poles. The hypotheses say that its only zero inside Γ is the simple zero a.

given
2.1

Apply [L1] to the meromorphic function fw and the holomorphic test function g(ζ)=ζ. By step 1.1, there is only one zero contribution, its multiplicity is 1, and the additional hypothesis n(Γ,a)=1 makes that contribution exactly a. The left-hand side is exactly the displayed contour integral.

step 1.1L1
3.1

Therefore the contour integral equals a. The role of [L2] is to identify the unique enclosed zero as the unique preimage of w.

step 1.1step 2.1L2

5 · Examples, counterexamples and false statements

None yet.

Sources