Several Complex Variables

Holomorphy, Continuation Phenomena,

and Automorphisms

Chinmay Mondal
CU Roll: C91/PMT/241017

Department of Pure Mathematics

University of Calcutta

Supervisor: Dr. Dhananjoy Mandal

[Link]. Fourth Semester Examination · 2026

Prologue

One variable versus several

In classical function theory three notions coincide:

complex dierentiability ⇐⇒ local power series ⇐⇒ Cauchy's integral formula.

Passing to Cn , n ≥ 2, this equivalence survives  but the topological intuition

of the plane does not.

ˆ No Riemann mapping theorem in Cn : the ball Bn and the polydisc Pn are

biholomorphically inequivalent.
ˆ Hartogs' phenomenon (1906): an isolated singularity of a holomorphic
function is impossible once n ≥ 2.

Three pillars: analytic foundations  continuation & singularities  rigidity of

automorphisms.

2/20
Notation

Conventions
Norms. Wirtinger operators.
 1/2
= 12 (∂x − i ∂y ),
 ∂
∥z ∥ = ∑j |zj |2 , |z | = maxj |zj |. ∂ zj j j

Multi-index. α ∈ Nn , ∂
= 12 (∂x + i ∂y ).
∂z̄j j j

αj
|α| = ∑ αj , α! = ∏ αj !, z
α = ∏ zj . Polydisc.

( , ) = { z : |zj − aj | < pj }.
P a p

A ⊂⊂ B denotes A compact in B . Polydiscs play the role in Cn that

discs play in C.

3/20
Holomorphic Functions on Cn

Denition and the CauchyRiemann system

Denition
f ∈ O(W ) if every a ∈ W has a neighbourhood on which f equals a
convergent power series in (z − a).

Equivalently: separately holomorphic in each zj , plus joint continuity.

Generalised CauchyRiemann equations
∂f
For f ∈ C 1 (W ): f ∈ O(W ) ⇐⇒ ≡ 0 on W , for j = 1, . . . , n.
∂z̄j

Example. sin(z1 + z22 ) ∈ O(C2 ); but |z2 |2 sin z1 ∈

/ O(C2 ), since ∂/∂z̄2 ̸= 0.

4/20
The Polydisc Cauchy Formula

The fundamental representation

Cauchy's formula on a polydisc
If f ∈ O(W ) and P (a, p ) ⊂ W , then for z ∈ P (a, p ),

1 f (ξ ) d ξ
Z Z
f (z ) =
(2π i )n
··· .
|ξ 1 −a1 |=p1 |ξ n −a |=p
n
(n
ξ 1 − z 1 ) · · · (ξ n − zn )

Proof. Induction on n: freeze ξ 1 , apply the inductive hypothesis in

z2, . . . , zn , then the one-variable formula in z1 ; Fubini justies the
interchange.
Dierentiated formula
f (ξ ) d ξ
D α f (z ) = α! (2π i )−n
Z
.
∏ j j − zj ) α + 1
( ξ j

5/20
Power Series, Estimates, and Identity

Rigidity begins here

Expansion
D α f (a )
f ∈ O(P (a, p )) =⇒ f (z ) = ∑α (z − a)α on P (a, p ).
α!
Cauchy estimates
|D α f (a)| ≤ M α! p −α , where M = supP (a,p ) |f |.

Identity theorem
W connected, f ∈ O(W ) vanishes on a nonempty open subset =⇒ f ≡ 0
on W .
Caveat. Vanishing on a curve need not propagate; the vanishing set must contain a

genuine open piece.

6/20
Open Mapping and Maximum Modulus

Qualitative consequences
Open mapping theorem
W connected, f ∈ O(W ) non-constant =⇒ f carries open sets to open sets.

Proof idea. Restrict to a complex line through two points of distinct

value; reduce to the one-variable case.
Maximum modulus principle
W bounded and connected, f ∈ O(W ) non-constant:
|f (z )| < supζ ∈∂W lim supz →ζ |f (z )| for all z ∈ W .

Corollary
An interior maximum of |f | forces f to be constant.

7/20
Three Convergence Theorems

Weierstrass · Montel · Vitali

Weierstrass
fν → f uniformly on compacta, fν ∈ O(W ) =⇒ f ∈ O(W ), and
D α fν → D α f likewise.
Montel
A locally uniformly bounded F ⊂ O(W ) is normal: every sequence has a
subsequence converging uniformly on compacta.
Vitali
(fν ) uniformly bounded, pointwise convergent on a set of uniqueness =⇒
convergence is uniform on compacta of W .

8/20
The Hartogs Phenomenon

A theorem with no one-variable analogue

In C, holomorphic functions may have natural boundaries . For n ≥ 2 this
is false.
Hartogs' extension theorem
n ≥ 2; P the unit polydisc; V ⊃ ∂P open, V ∩ P connected. Then every
f ∈ O(V ) extends to F ∈ O(P ∪ V ).

Mechanism. Laurent-expand f (z1 , z ′ ) = ∑ν aν (z ′ ) z1ν . On a thin shell,

aν ≡ 0 for ν < 0; since each aν ∈ O , the identity theorem propagates
this vanishing across all of parameter space.

Coecient-vanishing by analytic continuation  the engine of this chapter.

9/20
Reinhardt Domains

Laurent series in several variables

Denition
W is Reinhardt if z ∈ W =⇒ (e i θ1 z1 , . . . , e i θ zn ) ∈ W , ∀ θ ∈ Rn .
n

Theorem
On a Reinhardt domain, every f ∈ O(W ) has a unique Laurent expansion
∑α∈Z aα z α , uniform on compacta.
n

Corollaries.

ˆ Every coordinate vanishing somewhere =⇒ a genuine Taylor series

(α ∈ Nn ).
ˆ f extends to the lled-in Reinhardt hull W
c.

ˆ Holomorphy on a shell r < ∥z ∥ < R forces extension to the ball ∥z ∥ < R.

10/20
Singularities Are No Longer Isolated

The death of the isolated singularity

Theorem
For n ≥ 2: f ∈ O B \ {0} =⇒ f extends holomorphically across 0.

Consequence. The codimension-one substitute for a pole or essential

singularity is a singularity along an analytic hypersurface Z = {h = 0}.
f = z2 /z1 has a pole along {z1 = 0}; exp(1/z1 ) has an essential singularity
along {z1 = 0}.
Analytic sets
A ⊂ W is analytic if locally A = {f1 = · · · = fp = 0}, fi ∈ O . If A ⊊ W : A
closed, W \ A open, dense, connected.

11/20
Removability: Riemann, Radó, Hartogs

When the gap lls itself in

Riemann
A ⊂ W analytic, f ∈ O(W \ A) locally bounded near A =⇒ f extends to
O(W ).
Radó
f continuous on W , holomorphic on {f ̸= 0} =⇒ f ∈ O(W ).
Hartogs' continuity theorem
f ∈ O(W × Q ), Q an annulus; extension across one full disc at a ∈ W =⇒
extension to all of W × D.

12/20
Domains of Holomorphy

Pseudoconvexity and the Levi problem

Denition
W is a domain of holomorphy if no f ∈ O(W ) continues automatically
across an overlap into a strictly larger domain.

Polydiscs qualify; the Hartogs shell does not.

Plurisubharmonicity & pseudoconvexity
u is psh if λ 7→ u (a + λb ) is subharmonic on every complex line (log |f |
always is). W is pseudoconvex if − log dist(z , ∂W ) is psh.

Theorem (OkaBremermannNorguet)
W is a domain of holomorphy ⇐⇒ W is pseudoconvex.

13/20
Rigidity I: Cartan's Uniqueness Theorem

Bounded domains resist deformation

H. Cartan
D bounded, f : D → D holomorphic, f (a) = a, linear term of f at a equal to
the identity. Then f ≡ idD .

Proof idea. If f (z ) = z + Pm (z − a) + · · · with m ≥ 2, Pm ̸= 0, the

k-th iterate has degree-m coecient k Pm  unbounded as k → ∞,
contradicting the Cauchy estimates on bounded D .

Bounded domains permit no innitesimal deformation at a xed point.

14/20
Rigidity II: Linearity and Rank

Circular domains and constant rank

Denition
D is circular if z ∈ D , θ ∈ R =⇒ e i θ z ∈ D.
H. Cartan: linearity
D circular, 0 ∈ D, f ∈ Aut(D ) xes 0 =⇒ f (z ) = Az for invertible A.

Unitary maps on Bn , composed with Möbius-type maps moving the origin,

exhaust Aut(Bn ).
Rank theorem
f : W → Cm holomorphic of constant rank k =⇒ locally biholomorphic to
( z 1 , . . . , z k , 0, . . . , 0 ) .

15/20
Rigidity III: Osgood's Theorem

Injectivity forces biholomorphy

Theorem
W ⊂ Cn open, connected; f : W → Cn holomorphic and injective. Then
f (W ) is open and f −1 ∈ O(f (W )).

Sketch.
1. det(df ) ̸≡ 0, else the rank theorem violates injectivity.
2. f is locally biholomorphic o Λ = {det df = 0}.
3. f is proper on small polydiscs.
4. The local inverse extends across Λ via Radó's theorem.

Corollary: a bijective holomorphic self-map is automatically an automorphism.

16/20
Beyond Pure Mathematics

A survey of applications

ˆ Physics: CalabiYau manifolds, ˆ Robotics: the rank theorem

twistor theory locates kinematic singularities
ˆ Control theory: stability via ˆ Numerics: polydisc
singularities of transfer functions contour-integral methods
ˆ Imaging: analytic continuation in ˆ Learning theory: holomorphic
MRI reconstruction kernels (Fock, Bergman)

Unifying thread: holomorphy is far more rigid than smoothness  rigidity is

what makes continuation, compactness, and uniqueness so powerful.

17/20
Conclusion

What persists, and what is genuinely new

ˆ Persists from one variable: local power series, the polydisc Cauchy
formula, open mapping, maximum modulus, MontelVitali
compactness.
ˆ Genuinely new: no isolated singularities (Hartogs); pseudoconvexity
= domain of holomorphy (the Levi problem); injective =⇒
biholomorphic.

Beyond this dissertation: ∂¯ -cohomology and the full Levi problem; bounded
symmetric domains; Stein spaces and complex manifolds.

18/20
References

1. R. Narasimhan, Several Complex Variables, University of Chicago Press.

2. K. Fritzsche, H. Grauert, From Holomorphic Functions to Complex Manifolds,

GTM, Springer.

3. B. V. Shabat, Introduction to Complex Analysis, Part II, AMS.

4. L. Hörmander, An Introduction to Complex Analysis in Several Variables,

North-Holland.

5. S. G. Krantz, Function Theory of Several Complex Variables, AMS Chelsea.

6. R. M. Range, Holomorphic Functions and Integral Representations in Several

Complex Variables, GTM, Springer.

19/20
Thank You