06.09.08 · riemann-surfaces / stein

Survey: Cartan-Serre Stein theory in higher dim

shipped3 tiersLean: none

Anchor (Master): Stein 1951 *Analytische Funktionen mehrerer komplexer Veränderlichen zu vorgegebenen Periodizitätsmoduln und das zweite Cousinsche Problem* (Math. Ann. 123, originator: Stein-manifold definition); Cartan 1951–53 séminaire (originator: Theorems A and B); Cartan-Serre 1953 *Un théorème de finitude concernant les variétés analytiques compactes* (CRAS 237, originator: Theorem B in arbitrary complex dimension); Remmert 1956 *Sur les espaces analytiques holomorphiquement séparables et holomorphiquement convexes* (CRAS 243, originator: Remmert embedding); Bishop 1961 *Mappings of partially analytic spaces* (Amer. J. Math. 83, originator: sharp embedding dimension); Narasimhan 1960 *Imbedding of holomorphically complete complex spaces* (Amer. J. Math. 82, originator: sharp embedding for Stein spaces); Behnke-Stein 1949 *Entwicklung analytischer Funktionen auf Riemannschen Flächen* (Math. Ann. 120, originator: dimension-one ancestor); Hörmander *An Introduction to Complex Analysis in Several Variables* Ch. IV–V (textbook); Grauert-Remmert *Theory of Stein Spaces* (Grundlehren 236, canonical reference); Forstnerič *Stein Manifolds and Holomorphic Mappings* (2nd ed., Springer 2017, Oka principle); Gromov 1989 *Oka's principle for holomorphic sections of elliptic bundles* (J. Amer. Math. Soc. 2)

Intuition [Beginner]

A Stein manifold is the higher-dimensional cousin of a non-compact Riemann surface. In dimension one, every non-compact connected Riemann surface is Stein, by Behnke-Stein 06.09.03. In dimension two and higher, Stein-ness is a real restriction: most open subsets of $C^{n}$ are Stein, but not all open complex manifolds carry the holomorphic-function abundance the property demands.

The defining picture: a complex manifold $X$ is Stein when it carries a rich supply of global holomorphic functions. Rich enough to separate any two points (some holomorphic $f$ takes different values at $p$ and $q$ ). Rich enough to give every tangent direction its own complex coordinate. Rich enough to keep the holomorphic hull of every compact set compact (no run-away behaviour at infinity). These three conditions — separation, spreading, holomorphic convexity — are due to Stein 1951.

The pay-off of being Stein, in any complex dimension, is the Cartan-Serre theorem package: cohomology of every coherent analytic sheaf vanishes in positive degree, and global sections generate every stalk. Cousin problems, Mittag-Leffler, Runge approximation — every classical existence question collapses to one cohomological identity.

Visual [Beginner]

A schematic of a Stein manifold $X$ in complex dimension $n$ , drawn as a hill-shaped open complex submanifold of $C^{N}$ via the Bishop-Remmert embedding. Inside the hill, a strictly plurisubharmonic exhaustion function $φ : X \to R$ is sketched as concentric level sets ${φ = c}$ forming nested compact sublevel sets, each its own piece in the Stein exhaustion $X_{1} \subset X_{2} \subset X_{3} \subset \dots$ of $X$ . A separate panel contrasts: a polydisc and a convex bounded domain in $C^{2}$ (both Stein), versus the Hartogs shell ${1 < ∣ z ∣ < 2} \subset C^{2}$ (not Stein: every holomorphic function extends to the inner ball).

$Schematic placeholder for Stein manifolds in higher complex dimension: strictly plurisubharmonic exhaustion, proper embedding into $\mathbb{C}^N$, and the Hartogs shell as the canonical non-Stein.$

Worked example [Beginner]

Take $X = C^{n}$ — the simplest Stein manifold in dimension $n$ . The function $φ (z) = ∣ z ∣^{2} = ∣ z_{1} ∣^{2} + \dots + ∣ z_{n} ∣^{2}$ is a smooth real-valued function on $C^{n}$ that goes to infinity as $z \to \infty$ (exhaustion) and is strictly plurisubharmonic. Sublevel sets ${φ \leq c}$ are closed balls, compact. Holomorphic functions on $C^{n}$ separate points: for $p \neq = q$ there is a coordinate $z_{i}$ with $p_{i} \neq = q_{i}$ . The structure sheaf has no first cohomology: every Cauchy-Riemann-closed smooth $(0, 1)$ -form $g$ on $C^{n}$ can be solved as a Cauchy-Riemann image $g = D u$ for some smooth $u$ , by the Cauchy-Pompeiu formula in each variable.

Concrete instance: take $n = 2$ and the coherent sheaf $O_{C^{2}}$ . By Cartan's Theorem B, $H^{q} (C^{2}, O) = 0$ for every $q \geq 1$ . The Mittag-Leffler problem on $C^{2}$ — prescribe a Laurent tail at each point of a discrete set — is solvable. The Cousin I problem on $C^{2}$ — glue local meromorphic data with holomorphic transition differences — is solvable. The Cousin II problem — glue local divisor data with multiplicative holomorphic transitions — is solvable in the simply connected case, conditional on a topological obstruction in $H^{2} (C^{2}, Z) = 0$ .

What this tells us: $C^{n}$ is the model Stein manifold, where every classical existence problem is solvable in closed form. The Cartan-Serre package promotes the $C^{n}$ picture to every Stein manifold of every complex dimension, replacing closed-form computations by cohomological vanishing identities.

Check your understanding [Beginner]

Exercise (easy, multiple choice).

A complex manifold $X$ of complex dimension $n$ is called Stein when it satisfies which three conditions?

A. Compact, simply connected, Kähler B. Holomorphically convex, holomorphically separable, holomorphically spread C. Projective, smooth, of general type D. Compact, of finite genus, of finite type

Hint

The three Stein axioms are due to Stein 1951 and characterise abundance of global holomorphic functions: every compact set has a compact holomorphic hull; any two points can be told apart by some global $f$ ; and local coordinates can be written in terms of global $f$ 's.

Answer

B. Feedback-correct: a Stein manifold is holomorphically convex (compact holomorphic hulls), holomorphically separable (some global $f$ separates any two points), and holomorphically spread (some $n$ global $f$ 's give local coordinates at every point). Feedback-wrong: A names compact Kähler manifolds, the opposite extreme — these have very few global holomorphic functions; C names projective varieties, also compact; D names compact Riemann surfaces, again compact and the opposite of Stein.

Exercise (easy, true/false).

The Hartogs shell $Ω = {(z_{1}, z_{2}) \in C^{2} : 1 < ∣ z_{1} ∣^{2} + ∣ z_{2} ∣^{2} < 2}$ is a Stein manifold.

Hint

The Hartogs extension theorem 06.07.02 says every holomorphic function on $Ω$ extends to the full ball ${∣ z ∣^{2} < 2}$ . Does this match the abundance of holomorphic functions a Stein manifold should carry?

Answer

False. Feedback-correct: by the Hartogs extension theorem, every holomorphic function on $Ω$ extends to the inner ball ${∣ z ∣^{2} < 1}$ , so the algebra $O (Ω) = O (B (0, 2))$ is the same as on the full ball. The holomorphic hull of the compact set ${∣ z ∣^{2} = 1.5}$ contains the inner ball ${∣ z ∣^{2} \leq 1}$ , which is not compact in $Ω$ . Holomorphic convexity fails, and $Ω$ is not Stein. Feedback-wrong: the dimension-one analogue — annuli in $C$ — are Stein, because the Hartogs extension is a strictly higher-dimensional ( $n \geq 2$ ) phenomenon; carrying that intuition over to $C^{2}$ is a common slip.

Formal definition [Intermediate+]

Let $X$ be a connected complex manifold of complex dimension $n$ , with structure sheaf $O_{X}$ .

Definition (Stein manifold, Stein 1951). $X$ is Stein when the following three conditions hold.

(1) Holomorphically convex. For every compact $K \subset X$ the holomorphic hull

\hat{K}_{O (X)} = {x \in X : ∣ f (x) ∣ \leq K sup ∣ f ∣ for every f \in O (X)}

is compact in $X$ .

(2) Holomorphically separable. For every $p \neq = q$ in $X$ there is $f \in O (X)$ with $f (p) \neq = f (q)$ .

(3) Holomorphically spread. For every $p \in X$ there are $f_{1}, \dots, f_{n} \in O (X)$ whose differentials $d f_{1} (p), \dots, d f_{n} (p)$ are linearly independent over $C$ — equivalently, $(f_{1}, \dots, f_{n})$ are local holomorphic coordinates at $p$ .

Equivalent characterisations (Grauert 1958, Hörmander 1965, Narasimhan 1960). $X$ is Stein iff any of the following hold.

Strictly plurisubharmonic exhaustion. There is a smooth $φ : X \to R$ with $φ^{- 1} ((- \infty, c])$ compact for every $c$ and with $i \partial \overset{ˉ}{\partial} φ > 0$ pointwise as a $(1, 1)$ -form.
Theorem B vanishing. $H^{q} (X, F) = 0$ for every coherent analytic sheaf $F$ on $X$ and every $q \geq 1$ .
$\overset{ˉ}{\partial}$ -solvability. For every $\overset{ˉ}{\partial}$ -closed smooth $(0, q)$ -form $g$ on $X$ with $q \geq 1$ there is a smooth $(0, q - 1)$ -form $u$ with $\overset{ˉ}{\partial} u = g$ , with Hörmander's weighted $L^{2}$ -estimate quantifying the choice.
Cousin I and Cousin II. Both Cousin problems are unconditionally solvable on $X$ (with Cousin II conditioned by a $H^{2} (X, Z)$ -topological obstruction equivalent to the Picard-group statement $Pic (X) ≅ H^{2} (X, Z)$ ).

Theorem A (Cartan-Serre 1953). Let $X$ be a Stein manifold of complex dimension $n$ and $F$ a coherent analytic sheaf on $X$ . For every $x \in X$ the stalk $F_{x}$ is generated as an $O_{X, x}$ -module by the images of global sections. Equivalently, the natural evaluation map

Γ (X, F) \otimes_{Γ (X, O_{X})} O_{X, x} ⟶ F_{x}

is surjective for every $x \in X$ .

Theorem B (Cartan-Serre 1953). Let $X$ be a Stein manifold and $F$ a coherent analytic sheaf on $X$ . Then $H^{q} (X, F) = 0$ for every $q \geq 1$ .

Bishop-Remmert-Narasimhan embedding (Remmert 1956, Narasimhan 1960, Bishop 1961). A Stein manifold of complex dimension $n$ admits a proper holomorphic embedding $X ↪ C^{N}$ for $N = 2 n + 1$ . Eliashberg-Gromov 1992 sharpened the embedding dimension to $N = ⌊ 3 n /2 ⌋ + 1$ for $n \geq 2$ . Stein manifolds are exactly the closed complex submanifolds of $C^{N}$ (Remmert): one direction is the embedding theorem; the other is that a closed complex submanifold of $C^{N}$ inherits Stein-ness from $C^{N}$ via restriction of the squared-distance exhaustion.

Notation. $Γ (X, F)$ denotes global sections; $F_{x}$ the stalk at $x$ ; $H^{q} (X, F)$ the $q$ -th sheaf cohomology, computed via Čech with a Stein cover or via the Dolbeault fine resolution; $O_{X, x}$ the local ring of holomorphic germs at $x$ ; $i \partial \overset{ˉ}{\partial} φ$ the Levi form of $φ$ .

Counterexamples to common slips

Compactness obstructs Stein-ness. A connected compact complex manifold $X$ has $Γ (X, O_{X}) = C$ (constants only), so the holomorphic-separability axiom fails immediately. A compact Riemann surface, a projective curve, $P^{n}$ , and an abelian variety are all not Stein.
Hartogs phenomenon obstructs Stein-ness in dimension $\geq 2$ . The Hartogs shell, the punctured ball $B ∖ {0}$ in $C^{n}$ for $n \geq 2$ , and more generally any open $U \subset C^{n}$ on which Hartogs extension forces a strict enlargement of $O (U)$ , fails Stein convexity. The dimension-one analogue (annuli, punctured discs) is Stein — Hartogs is a real obstruction only starting in $C^{2}$ .
Coherence is required for Theorem B. The vanishing $H^{q} (X, F) = 0$ holds only for coherent $F$ . The constant sheaf $\underline{Z}_{X}$ has non-vanishing higher cohomology determined by the topology of $X$ (e.g., $H^{1} (C^{*}, \underline{Z}) = Z$ ), independent of the Stein hypothesis.
Theorem A is module generation, not section equality. The evaluation map $Γ (X, F) \to F_{x}$ need not be surjective on its own; the surjection is the evaluation tensored with the local ring $O_{X, x}$ . Local coefficients are part of the statement.

Key theorem with proof [Intermediate+]

Theorem (Cartan-Serre 1953, Theorem B in arbitrary complex dimension). Let $X$ be a Stein manifold of complex dimension $n$ and $F$ a coherent analytic sheaf on $X$ . Then $H^{q} (X, F) = 0$ for every $q \geq 1$ .

Proof. The argument runs in five steps: a strictly plurisubharmonic exhaustion, the Hörmander $L^{2}$ -existence theorem on each sublevel set, propagation to coherent sheaves via syzygies, a Cartan-Serre limit passage to the union, and the dimension-induction step that handles general $q$ .

Step 1 — Stein exhaustion by strictly plurisubharmonic sublevel sets. By the equivalent characterisation, $X$ carries a smooth strictly plurisubharmonic exhaustion $φ : X \to R$ . Choose regular values $c_{1} < c_{2} < \dots$ with $c_{n} \to \infty$ and define $X_{n} = {φ < c_{n}}$ . Each $X_{n}$ is a relatively compact open with smooth strictly pseudoconvex boundary, hence itself a Stein manifold (the restriction of $φ$ is a strictly plurisubharmonic exhaustion of $X_{n}$ with finite supremum). The Stein opens form an increasing exhaustion $X_{1} ⋐ X_{2} ⋐ \dots$ with $⋃_{n} X_{n} = X$ .

Step 2 — Hörmander $L^{2}$ -solvability of $\overset{ˉ}{\partial}$ on each $X_{n}$ . Hörmander's weighted $L^{2}$ -existence theorem (Hörmander 1965 Acta Math. 113) states the following. On a Stein manifold $Y$ of complex dimension $n$ with strictly plurisubharmonic exhaustion $φ$ , for every $\overset{ˉ}{\partial}$ -closed smooth $(0, q)$ -form $g$ on $Y$ with $q \geq 1$ and every weight function $ϕ$ of sufficiently positive Levi form, there is a smooth $(0, q - 1)$ -form $u$ on $Y$ with $\overset{ˉ}{\partial} u = g$ and $∥ u ∥_{L^{2} (ϕ)} \leq C ∥ g ∥_{L^{2} (ϕ)}$ . Specialised to $Y = X_{n}$ and $ϕ = K φ$ for $K$ large, the theorem solves $\overset{ˉ}{\partial} u = g$ for every $\overset{ˉ}{\partial}$ -closed $g$ on $X_{n}$ . By the Dolbeault fine-resolution comparison, $H^{q} (X_{n}, O_{X_{n}}) = 0$ for every $q \geq 1$ .

Step 3 — propagation to coherent $F$ on $X_{n}$ . Oka's coherence theorem (Oka 1950) makes $O_{X}$ coherent as a sheaf over itself. Locally on $X_{n}$ , a coherent sheaf $F$ admits finite presentations $O^{n_{1}} \to O^{n_{0}} \to F \to 0$ . Iterating the kernel-of-presentation construction on the relatively compact Stein open $X_{n}$ and applying Hilbert's syzygy theorem to the regular local rings $O_{X, x}$ of Krull dimension $n$ — these are regular of global homological dimension $n$ — produces a finite free resolution $0 \to O^{m_{n}} \to \dots \to O^{m_{0}} \to F \to 0$ of length at most $n$ on $X_{n}$ . The long exact sequence in cohomology applied to this resolution, combined with the Step 2 vanishing $H^{q} (X_{n}, O_{X_{n}}) = 0$ for $q \geq 1$ , forces $H^{q} (X_{n}, F) = 0$ for $q \geq 1$ by an induction on the resolution length. (The dimension-one specialisation in 06.09.02 uses resolution length one, the syzygy collapsing immediately because $dim X = 1$ ; the higher-dimensional case requires the full induction.)

Step 4 — limit passage from $X_{n}$ to $X$ . The full manifold $X = ⋃_{n} X_{n}$ is the increasing union of Stein opens. The Čech complex of $F$ on $X$ relative to a refinement of Stein covers of each $X_{n}$ is the inverse limit (in cochain degree) of the Čech complexes on $X_{n}$ . The Mittag-Leffler condition on the inverse system ${Γ (X_{n}, F)}$ — each restriction $Γ (X_{n + 1}, F) \to Γ (X_{n}, F)$ has dense image in the natural Fréchet topology, by Cartan's Runge-type approximation theorem on Stein subdomains — allows the inverse limit to commute with cohomology in degree one. Concretely: a class $[ξ] \in H^{q} (X, F)$ restricts to classes $[ξ_{n}] \in H^{q} (X_{n}, F)$ , each vanishing by Step 3, with local primitives $η_{n}$ on $X_{n}$ . The density condition produces a global primitive $η$ on $X$ by uniformly approximating each correction term $η_{n + 1} - η_{n}$ on $X_{n - 1}$ and assembling the geometric sum. Therefore $H^{q} (X, F) = 0$ for $q \geq 1$ .

Step 5 — induction on $q$ via Hilbert's syzygy. The argument of Steps 1–4 reduces vanishing of $H^{q}$ to vanishing of $H^{q}$ on each Stein sublevel set, and the resolution-length induction within Step 3 handles the propagation from $O_{X}$ to arbitrary coherent $F$ . Because the resolution length is finite (bounded by $n = dim_{C} X$ ), every $H^{q} (X, F)$ with $q \geq 1$ is built from $H^{q} (X_{n}, O_{X_{n}})$ -vanishings, all of which vanish in Step 2. The cohomology vanishes in every positive degree. $□$

Theorem A follows from Theorem B by a one-step exact-sequence argument: given $σ \in F_{x}$ , extend to a local section $\tilde{σ}$ on a neighbourhood $U$ of $x$ , form the sheaf morphism $O_{X} \to F$ defined by $1 \mapsto \tilde{σ}$ on $U$ and zero outside, with kernel $K$ coherent; the long exact sequence with $H^{1} (X, K) = 0$ (Theorem B) forces surjectivity of the evaluation map on global sections at $x$ .

Bridge. The proof binds two pieces of upstream machinery into one statement. From the dimension-one ancestor 06.09.02 comes the Stein-exhaustion plus Hörmander- $L^{2}$ plus Mittag-Leffler-limit template, with the syzygy collapsing to length one in real dimension two. From the higher-dimensional Hörmander 1965 theorem comes the weighted $L^{2}$ -existence for $\overset{ˉ}{\partial}$ on a strictly plurisubharmonic exhaustion of arbitrary complex dimension. The Cartan-Serre 1953 paper extracted this chain in arbitrary complex dimension; the dimension-one version in Forster §29 is the cleanest expression because the syzygy collapses to one step, while the higher-dimensional case builds toward the elliptic-PDE proof via Hörmander's $L^{2}$ -method. The foundational reason every classical existence problem on a Stein manifold reduces to Theorem B is exactly the cohomological packaging: Cousin I 06.09.04 is the $H^{1} (X, O_{X})$ -obstruction, Cousin II 06.09.05 is the $H^{1} (X, O_{X}^{*})$ -obstruction, Mittag-Leffler 06.09.06 is the principal-parts $δ$ -connecting map into $H^{1} (X, O_{X})$ , and Runge approximation 06.09.07 is the surjectivity of compactly-supported $H_{c}^{1}$ restrictions dual to the $H^{0}$ -density statement. Putting these together, the central insight is that the Cartan-Serre vanishing identifies the Stein hypothesis with the universal solvability of every coherent-sheaf cohomology problem, and the bridge is the Hörmander weighted $L^{2}$ -estimate that makes this identification quantitative. The same chain appears again in the Oka principle on principal bundles and in the Forstnerič-Gromov elliptic h-principle.

Exercises [Intermediate+]

Exercise 3 (medium, symbolic).

Let $X$ be a Stein manifold of complex dimension $n$ and $D = D_{1} + \dots + D_{k}$ a divisor on $X$ with $D_{i}$ smooth analytic hypersurfaces meeting transversally. The ideal sheaf $I_{D}$ of $D$ is coherent. Show that the sequence

0 \to I_{D} \to O_{X} \to O_{D} \to 0

induces an isomorphism $Γ (X, O_{X}) /Γ (X, I_{D}) ≅ Γ (X, O_{D})$ .

Hint

Use the long exact sequence in cohomology and Theorem B applied to the coherent sheaves $I_{D}$ and $O_{X}$ .

Answer

The short exact sequence is the ideal-of-divisor sequence: $O_{D} = O_{X} / I_{D}$ is the structure sheaf of the divisor as an analytic subspace. The long exact sequence in cohomology reads

0 \to Γ (X, I_{D}) \to Γ (X, O_{X}) \to Γ (X, O_{D}) \to H^{1} (X, I_{D}) \to H^{1} (X, O_{X}) \to \dots .

By Cartan-Serre Theorem B applied to the coherent sheaves $I_{D}$ and $O_{X}$ on the Stein $X$ , both $H^{1}$ -terms vanish. The connecting map's target therefore vanishes, and the four-term sequence reduces to a short exact

0 \to Γ (X, I_{D}) \to Γ (X, O_{X}) \to Γ (X, O_{D}) \to 0,

giving the desired quotient identification. Rubric: full credit for the long exact sequence + double application of Theorem B + identification with $Γ (X, O_{D})$ .

Exercise 4 (medium, short-answer).

Show that the Cousin I problem on a Stein manifold $X$ — given an open cover ${U_{i}}$ and meromorphic functions $f_{i} \in M (U_{i})$ with holomorphic differences $f_{i} - f_{j} \in O (U_{i} \cap U_{j})$ — is always solvable.

Hint

The obstruction is the cohomology class $[f_{i} - f_{j}] \in H^{1} (X, O_{X})$ .

Answer

The differences $g_{ij} = f_{i} - f_{j}$ are holomorphic on $U_{ij}$ by hypothesis and satisfy the Čech cocycle condition $g_{ij} + g_{j k} + g_{k i} = 0$ on triple overlaps automatically. The cocycle defines a class $[g] \in H^{1} (X, O_{X})$ . By Cartan-Serre Theorem B applied to the coherent sheaf $O_{X}$ on the Stein $X$ , $H^{1} (X, O_{X}) = 0$ , so the cocycle is a coboundary $g_{ij} = h_{i} - h_{j}$ for some $h_{i} \in O (U_{i})$ . The function $f$ defined locally by $f = f_{i} - h_{i}$ on $U_{i}$ is well-defined globally (because $f_{i} - h_{i} = f_{j} - h_{j}$ on $U_{ij}$ by the coboundary identity) and meromorphic on $X$ with $f - f_{i} \in O (U_{i})$ . Rubric: full credit for the cocycle + Theorem B + assembly of global $f$ .

Exercise 5 (medium, short-answer).

State the Bishop-Remmert-Narasimhan embedding theorem and explain in what sense Stein manifolds are exactly the closed complex submanifolds of $C^{N}$ .

Hint

One direction is the embedding theorem itself; the other uses the restriction of the squared-distance exhaustion of $C^{N}$ to a closed submanifold.

Answer

Bishop-Remmert-Narasimhan embedding theorem. A complex manifold $X$ of complex dimension $n$ is Stein if and only if it admits a proper holomorphic embedding $X ↪ C^{N}$ for some $N$ (and the sharp embedding dimension is $N = 2 n + 1$ by Remmert 1956 / Narasimhan 1960, refined to $N = ⌊ 3 n /2 ⌋ + 1$ by Eliashberg-Gromov 1992 for $n \geq 2$ ).

Direction 1 (Stein → embedding): the three Stein axioms produce enough global holomorphic functions to separate points (axiom 2) and spread coordinates (axiom 3); together with holomorphic convexity (axiom 1), one constructs a proper embedding using $2 n + 1$ carefully chosen $f_{i} \in O (X)$ . The construction uses a Whitney-style transversality argument on the holomorphic-convex hull data to ensure properness and embedding-immersion.

Direction 2 (closed submanifold of $C^{N}$ → Stein): let $X ↪ C^{N}$ be a closed complex submanifold. The function $φ (x) = ∣ x ∣^{2}$ on $C^{N}$ restricts to a strictly plurisubharmonic function on $X$ (the Levi form is the pull-back of the standard positive Levi form on $C^{N}$ via the embedding). The closedness of $X$ in $C^{N}$ makes $φ ∣_{X}$ an exhaustion. The strictly plurisubharmonic exhaustion characterisation of Stein-ness then gives Stein.

The two directions together identify Stein manifolds with closed complex submanifolds of $C^{N}$ — Stein-ness is the abstract geometric face of being embedded in complex affine space. Rubric: full credit for both directions and the dimension bounds.

Exercise 6 (medium, short-answer).

Explain why the Cousin II problem on a Stein manifold $X$ — given a cover ${U_{i}}$ and non-zero meromorphic $f_{i} \in M^{*} (U_{i})$ with holomorphic non-vanishing transition ratios $f_{i} / f_{j} \in O^{*} (U_{ij})$ — is conditionally solvable, with obstruction in $H^{2} (X, Z)$ .

Hint

The multiplicative cocycle $g_{ij} = f_{i} / f_{j}$ defines a class in $H^{1} (X, O^{*}) = Pic (X)$ ; the exponential exact sequence relates this to $H^{1} (X, O)$ and $H^{2} (X, Z)$ .

Answer

The ratios $g_{ij} = f_{i} / f_{j}$ define a Čech 1-cocycle with values in $O_{X}^{*}$ , and the class $[g] \in H^{1} (X, O_{X}^{*}) = Pic (X)$ is the Cousin II obstruction. The exponential exact sequence of sheaves $0 \to Z \to O_{X} \to O_{X}^{*} \to 0$ induces the long exact sequence in cohomology

\dots \to H^{1} (X, O_{X}) \to H^{1} (X, O_{X}^{*}) \to H^{2} (X, Z) \to H^{2} (X, O_{X}) \to \dots .

By Cartan-Serre Theorem B applied to the coherent sheaf $O_{X}$ on the Stein $X$ , both $H^{1} (X, O_{X}) = 0$ and $H^{2} (X, O_{X}) = 0$ . The exact sequence collapses to an isomorphism

Pic (X) = H^{1} (X, O_{X}^{*}) ≅ H^{2} (X, Z) .

The Cousin II problem is therefore solvable exactly when the topological Chern class $c_{1} \in H^{2} (X, Z)$ of the associated divisor data vanishes. On a topologically simple Stein $X$ (e.g., $C^{n}$ for any $n \geq 1$ , or a contractible Stein), $H^{2} (X, Z) = 0$ and Cousin II is unconditionally solvable; on a topologically non-simple Stein $X$ (e.g., $C^{*} \times C^{*} \subset C^{2}$ , with $H^{2} (X, Z) = Z$ ), the obstruction is real and Cousin II can fail. Rubric: full credit for the exponential exact sequence + the $Pic (X) ≅ H^{2} (X, Z)$ identification + the obstruction discussion.

Exercise 7 (hard, short-answer).

State the Levi problem and explain how its resolution (Bremermann / Norguet / Oka 1953-54) identifies pseudoconvex domains in $C^{n}$ with Stein manifolds.

Hint

A domain of holomorphy is an open $Ω \subset C^{n}$ such that no proper extension across the boundary exists; pseudoconvexity is the local condition that the boundary's Levi form is positive semidefinite.

Answer

Levi problem (E. E. Levi 1910, posed; Oka 1942 / Bremermann 1954 / Norguet 1954 / Oka 1953 solved). Let $Ω \subset C^{n}$ be an open set. Three a priori distinct convexity conditions are equivalent:

Domain of holomorphy: there is some $f \in O (Ω)$ which has no holomorphic extension across the boundary $\partial Ω$ ; equivalently, the holomorphic functions on $Ω$ separate points of $\partial Ω$ in the limit.
Pseudoconvex: $- lo g d (\cdot, \partial Ω)$ is plurisubharmonic on $Ω$ (equivalently, $Ω$ admits a strictly plurisubharmonic exhaustion).
Stein: $Ω$ is holomorphically convex, holomorphically separable, holomorphically spread.

Proof of equivalence (sketch). $(1) \Rightarrow (2)$ : a domain of holomorphy has its envelope-of-holomorphy radius coincident with the boundary distance; the logarithm of a holomorphy radius is plurisubharmonic by classical $(0, 1)$ -form integration. $(2) \Rightarrow (3)$ : from the strictly plurisubharmonic exhaustion, the equivalent characterisation in the Formal definition gives Stein-ness; this is the analytically deepest step, requiring Hörmander's weighted $L^{2}$ -existence theorem or the original Oka / Bremermann / Norguet exhaustion construction. $(3) \Rightarrow (1)$ : a Stein manifold admits global holomorphic functions separating points up to and including the boundary, by the Bishop-Remmert embedding; pulled back to $Ω$ , these produce non-extendable holomorphic functions.

The equivalence identifies the complex-analytic notion of Stein-ness with the real-PDE notion of pseudoconvexity (positive Levi form) and the function-theoretic notion of domain of holomorphy. It is the higher-dimensional generalisation of the dimension-one statement that every non-compact Riemann surface is automatically Stein — in dimension one, every open set is a domain of holomorphy (annuli, punctured discs, half-planes), and the Levi-problem equivalence becomes vacuous. In dimension $n \geq 2$ , the Levi equivalence is a substantive PDE-and-function-theory theorem; the Hartogs phenomenon 06.07.02 is the obstruction to the equivalence when the pseudoconvexity hypothesis fails. Rubric: full credit for stating the three-way equivalence and naming the original proofs.

Exercise 8 (hard, short-answer).

State the Oka principle in its Grauert form (Grauert 1958) and explain why it is the higher-dimensional generalisation of the dimension-one Cousin II solvability on a non-compact Riemann surface.

Hint

Oka's principle says that on a Stein source, holomorphic and continuous classification problems for principal bundles with complex Lie group fibre coincide.

Answer

Oka principle (Oka 1939 / Grauert 1958). Let $X$ be a Stein manifold and $G$ a complex Lie group. The natural map of classifying sets

{holomorphic principal G -bundles on X} / ≅ ⟶ {continuous principal G -bundles on X} / ≅

is a bijection. Equivalently, every continuous principal $G$ -bundle on a Stein manifold is isomorphic, by a continuous gauge transformation, to a holomorphic principal $G$ -bundle, and the isomorphism class is determined by the underlying topological class.

Dimension-one specialisation. For $G = C^{*}$ (the multiplicative group), a principal $G$ -bundle is the same as a holomorphic line bundle, and the classifying set is $Pic (X) = H^{1} (X, O^{*})$ . By the exponential sequence (Exercise 6), $Pic (X) ≅ H^{2} (X, Z)$ on a Stein manifold. On a non-compact Riemann surface, $H^{2} (X, Z) = 0$ for topological reasons (non-compact real-2-manifold), so $Pic (X) = 0$ — every line bundle is holomorphically isomorphic to the structure sheaf. This is the Cousin II solvability on a non-compact RS 06.09.05.

Higher-dimensional content. In complex dimension $n \geq 2$ , the topological cohomology $H^{2} (X, Z)$ on a Stein $X$ is genuinely interesting (e.g., $H^{2} (C^{*} \times C^{*}, Z) = Z$ ), and the Oka principle identifies the holomorphic Picard group with this topological invariant. The Grauert 1958 paper extends the statement to every complex Lie group fibre, including $GL_{n} (C)$ , $SL_{n} (C)$ , and discrete groups. The Forstnerič-Gromov elliptic h-principle (Gromov 1989, Forstnerič 2003-) extends to holomorphic maps Stein → elliptic target, with the homotopy classification of continuous maps equivalent to the holomorphic classification. Rubric: full credit for the Grauert statement + the dimension-one specialisation to Cousin II + the explanation of why the higher-dimensional case has real content.

Lean formalization [Intermediate+]

Mathlib does not currently formalise the Cartan-Serre Stein theory in higher complex dimension. A proposed signature, in Lean 4 / Mathlib syntax, sketching the target theorems:

-- Sketch only; no current Mathlib coverage. See lean_mathlib_gap.
import Mathlib.Geometry.Manifold.ComplexManifold
import Mathlib.Analysis.Complex.Polydisc
import Mathlib.AlgebraicGeometry.Sheaf
import Mathlib.CategoryTheory.Abelian.DerivedFunctor

namespace Codex.RiemannSurfaces.Stein.Survey

variable (n : ℕ)
variable (X : Type*) [ComplexManifold X] [FiniteDimensional ℂ X]

-- A Stein manifold: holomorphically convex, holomorphically separable,
-- holomorphically spread. Three predicates packaged as one.
structure SteinManifold (X : Type*) [ComplexManifold X] : Prop where
  holConvex : ∀ K : Set X, IsCompact K → IsCompact (holHull K)
  holSeparable : ∀ p q : X, p ≠ q → ∃ f : X → ℂ, IsHolomorphic f ∧ f p ≠ f q
  holSpread : ∀ p : X, ∃ f : Fin (finrank ℂ X) → (X → ℂ),
                (∀ i, IsHolomorphic (f i)) ∧
                LinearIndependent ℂ (fun i ↦ (mfderiv ℂ ℂ (f i) p) 1)

variable (hX : SteinManifold X)
variable (F : CoherentAnalyticSheaf X)

-- Cartan-Serre Theorem A: global sections generate stalks.
theorem cartanSerre_theoremA (x : X) :
    Function.Surjective
      (fun (s : Γ X F ⊗[Γ X (𝓞_X)] (𝓞_X.stalk x)) ↦ s.eval x) := by
  sorry

-- Cartan-Serre Theorem B: higher cohomology vanishes.
theorem cartanSerre_theoremB {q : ℕ} (hq : 1 ≤ q) :
    Sheaf.cohomology ℂ q X F.toSheaf = ⊥ := by
  sorry

-- Bishop-Remmert-Narasimhan embedding: X embeds properly into ℂ^N.
theorem bishopRemmertNarasimhan_embedding :
    ∃ (N : ℕ), ∃ (φ : X → (Fin N → ℂ)),
      IsHolomorphic φ ∧ IsClosedEmbedding φ ∧ N ≤ 2 * (finrank ℂ X) + 1 := by
  sorry

end Codex.RiemannSurfaces.Stein.Survey

The proof depends on names absent from Mathlib (the coherent-analytic-sheaf category on a complex manifold of arbitrary dimension; Oka's coherence theorem; Hörmander's weighted $L^{2}$ -existence for $\overset{ˉ}{\partial}$ on a strictly plurisubharmonic exhaustion; the Cartan-Serre limit-passage; the Remmert / Narasimhan / Bishop embedding construction). Each is a candidate Mathlib contribution; until then this unit ships with lean_status: none.

Advanced results [Master]

The Cartan-Serre Stein theory is the higher-dimensional generalisation of the Behnke-Stein 1949 dimension-one theorem. The package — Theorems A and B, Bishop-Remmert embedding, Cousin solvability, Oka principle — collects every classical existence theorem on a non-compact complex manifold into a single cohomological identity. The survey gathers the higher-dimensional refinements that are absent from the dimension-one specialisation 06.09.02 and explains how each is the engine of a separate strand of modern complex geometry.

Stein manifold definition (Stein 1951). Karl Stein's 1951 paper Analytische Funktionen mehrerer komplexer Veränderlichen zu vorgegebenen Periodizitätsmoduln und das zweite Cousinsche Problem (Math. Ann. 123) introduced the axiomatic definition. Three conditions: holomorphic convexity (every compact $K$ has compact $O (X)$ -hull), holomorphic separability (any two distinct points are separated by some global $f$ ), and holomorphic spreading (some $n$ global $f$ 's give local coordinates at every point). Stein's motivation was the additive period-prescription problem in $C^{n}$ : given prescribed periodicity data on a covering of $C^{n}$ minus a discrete set, find a global meromorphic function realising the data. The three axioms isolate the abstract geometric condition that makes the problem solvable for every coherent-sheaf datum, in every complex dimension simultaneously.

Cartan-Serre Theorems A and B (1953). Henri Cartan's 1951–53 séminaire at the École Normale Supérieure and the Cartan-Serre 1953 Comptes Rendus note (CRAS 237, 128–130) established the cohomological theorems on Stein manifolds in arbitrary complex dimension. Theorem A: global sections generate every stalk of every coherent analytic sheaf. Theorem B: $H^{q} (X, F) = 0$ for every $q \geq 1$ and every coherent analytic sheaf $F$ on a Stein manifold. The two theorems are equivalent (Theorem A follows from Theorem B by a one-step exact sequence). The proof in the dimension-one case was given by Behnke-Stein 1949 06.09.03; the dimension- $n$ case in the Cartan séminaire used a syzygy induction on the resolution length, combined with the local $L^{2}$ -existence theorem for $\overset{ˉ}{\partial}$ on a strictly plurisubharmonic exhaustion (which Hörmander 1965 supplied in its modern form).

Bishop-Remmert-Narasimhan embedding theorem. Reinhold Remmert in 1956 (C. R. Acad. Sci. Paris 243, 118–121) proved that every Stein manifold of complex dimension $n$ embeds properly and holomorphically into $C^{2 n + 1}$ . Raghavan Narasimhan in 1960 (Amer. J. Math. 82, 917–934) extended the result to Stein analytic spaces with singularities, and Errett Bishop in 1961 (Amer. J. Math. 83, 209–242) refined the embedding construction. Stein manifolds are exactly the closed complex submanifolds of $C^{N}$ — one direction is the embedding theorem, the other is restriction of the squared-distance exhaustion. Eliashberg-Gromov 1992 (Ann. of Math. 136, 123–135) sharpened the embedding dimension to $N = ⌊ 3 n /2 ⌋ + 1$ , which is optimal in general (sharp for $n \geq 2$ ).

The Cousin problems in higher dimension. Pierre Cousin's 1895 paper Sur les fonctions de $n$ variables complexes (Acta Math. 19, 1–62) posed the two Cousin problems on bidiscs in $C^{n}$ . Cousin I (additive): given an open cover ${U_{i}}$ and meromorphic $f_{i} \in M (U_{i})$ with $f_{i} - f_{j} \in O (U_{ij})$ holomorphic, find a global meromorphic $f$ with $f - f_{i} \in O (U_{i})$ . Cousin II (multiplicative): given a cover and non-zero meromorphic $f_{i}$ with $f_{i} / f_{j} \in O^{*} (U_{ij})$ , find a global $f$ with $f / f_{i} \in O^{*} (U_{i})$ . On a Stein manifold $X$ , Cousin I is unconditionally solvable: the obstruction $[f_{i} - f_{j}] \in H^{1} (X, O_{X}) = 0$ vanishes by Theorem B. Cousin II is conditionally solvable: the obstruction $[f_{i} / f_{j}] \in H^{1} (X, O_{X}^{*}) = Pic (X) ≅ H^{2} (X, Z)$ (via the exponential sequence on a Stein, with $H^{1} (X, O_{X}) = H^{2} (X, O_{X}) = 0$ ) reduces the solvability question to the topological Chern class.

Mittag-Leffler and Runge on Stein manifolds. The Mittag-Leffler theorem on a Stein manifold $X$ — prescribed Laurent tails along a discrete locally finite subset realised by a global meromorphic function — is the principal-parts specialisation of Cousin I, solvable by Theorem B. The Oka-Weil approximation theorem (Oka 1937; Weil 1935) — every holomorphic function on a polynomially convex compact set $K \subset X$ is uniformly approximable by global $O (X)$ -restrictions — is the Stein generalisation of the dimension-one Runge theorem 06.09.07 and follows from Theorem B applied to the kernel sheaf of the restriction map. Hörmander's textbook Several Complex Variables (Ch. III) packages both as one-page corollaries of $\overset{ˉ}{\partial}$ -solvability.

The Hartogs phenomenon and the Levi problem. The Hartogs extension phenomenon (Hartogs 1906, Math. Ann. 62, 1–88) is the failure of Stein-ness for non-pseudoconvex open subsets of $C^{n}$ with $n \geq 2$ : on a Hartogs shell ${1 < ∣ z ∣ < 2} \subset C^{2}$ , every holomorphic function extends to the full ball, so holomorphic convexity fails. The Levi problem (E. E. Levi 1910, posed) asks the converse: is every pseudoconvex domain a domain of holomorphy (and hence Stein)? Oka 1942 (Tôhoku Math. J. 49) solved the dimension-two case; the dimension- $n$ case was solved independently in 1953–54 by Oka, Bremermann, and Norguet. The Levi equivalence identifies three a priori distinct convexity conditions — domain of holomorphy, pseudoconvex, Stein — on open subsets of $C^{n}$ . Combined with Hörmander 1965, the entire Stein theory in $C^{n}$ rests on the strictly plurisubharmonic exhaustion as the load-bearing geometric input.

Oka principle. Kiyoshi Oka's 1939 paper Sur les fonctions analytiques de plusieurs variables III: Deuxième problème de Cousin (J. Sci. Hiroshima Univ. Ser. A 9) proved the dimension- $n$ Cousin II solvability and noted the Oka principle: holomorphic and continuous classifications of certain bundles on Stein sources coincide. Hans Grauert's 1958 paper Analytische Faserungen über holomorph-vollständigen Räumen (Math. Ann. 135, 263–273) generalised to principal $G$ -bundles for $G$ any complex Lie group. Mikhail Gromov's 1989 paper Oka's principle for holomorphic sections of elliptic bundles (J. Amer. Math. Soc. 2, 851–897) reformulated the principle as an h-principle: holomorphic maps Stein → elliptic target obey the same homotopy classification as continuous maps. Forstnerič and Larusson refined this into modern Oka theory (Forstnerič 2017 Stein Manifolds and Holomorphic Mappings 2nd ed.). The Oka principle is the higher-dimensional generalisation of the dimension-one Cousin II solvability and is the central engine of contemporary holomorphic homotopy theory.

Worked example (the polydisc). $Δ^{n} = {(z_{1}, \dots, z_{n}) : ∣ z_{i} ∣ < 1} \subset C^{n}$ is the canonical bounded Stein domain. The function $φ (z) = \sum_{i} - lo g (1 - ∣ z_{i} ∣^{2})$ is a strictly plurisubharmonic exhaustion of $Δ^{n}$ ; the Levi form is positive definite at every point because the second derivative of $- lo g (1 - ∣ w ∣^{2})$ at $∣ w ∣ < 1$ is positive. Cartan-Serre Theorem B on $Δ^{n}$ : every coherent analytic sheaf has vanishing higher cohomology. Cousin I and Cousin II are unconditionally solvable on $Δ^{n}$ (the polydisc is contractible, so $H^{2} (Δ^{n}, Z) = 0$ ). Runge approximation on $Δ^{n}$ is the multivariate Cauchy-Hadamard polynomial-approximation theorem: every holomorphic function on $Δ^{n}$ is uniformly approximable on compact subsets by polynomials in $z_{1}, \dots, z_{n}$ . The same picture extends to every bounded convex domain in $C^{n}$ via convexity-implies-pseudoconvexity-implies-Stein. Hartogs' theorem fails for the polydisc (it is genuinely an open subset of $C^{n}$ with non-extendable holomorphic functions, by maximum modulus on the distinguished boundary), confirming that boundedness and Stein-ness can coexist.

Worked example (a tube domain). Let $Ω = T_{C} = {z \in C^{n} : Im z \in C}$ for $C \subset R^{n}$ a convex open cone. The tube $T_{C}$ is a Stein manifold by Bochner's tube theorem: an open tube over a convex cone is a domain of holomorphy in $C^{n}$ . The proof: the function $φ (z) = - lo g d (Im z, \partial C)$ is plurisubharmonic and exhausts $T_{C}$ ; convexity of $C$ gives the strict plurisubharmonicity of $φ$ in the imaginary directions, and the Cauchy-Riemann equations transfer this to a strictly plurisubharmonic exhaustion. Tube domains over convex cones are the canonical examples of Stein manifolds with rich automorphism groups (Siegel upper half-space is a generalisation), and they appear in the boundary-value-problem theory of holomorphic functions and in representations of semisimple Lie groups (Bargmann, Gelfand-Naimark, Harish-Chandra).

Failure mode: the punctured ball. Let $B ∖ {0} \subset C^{n}$ for $n \geq 2$ . Every holomorphic function on $B ∖ {0}$ extends to $B$ by the Hartogs extension theorem 06.07.02; the function ring $O (B ∖ {0}) = O (B)$ is no bigger than on the full ball. Holomorphic convexity fails: the compact sphere ${∣ z ∣ = 1/2}$ has $O (B ∖ {0})$ -hull equal to the closed ball ${∣ z ∣ \leq 1/2}$ minus zero, which is not compact in $B ∖ {0}$ . So $B ∖ {0}$ is not Stein for $n \geq 2$ . By contrast, the punctured disc $Δ ∖ {0} \subset C$ is Stein (it is a non-compact Riemann surface, hence Stein by Behnke-Stein). The Hartogs obstruction is a genuinely higher-dimensional phenomenon.

Synthesis. The Cartan-Serre theory unifies every classical existence problem on a non-compact complex manifold into a single cohomological identity: Theorem B vanishing of $H^{q} (X, F)$ for $q \geq 1$ and every coherent $F$ . The foundational reason every Cousin, Mittag-Leffler, Runge, and bundle-classification problem reduces to one statement is exactly the coherence of the relevant ideal / quotient sheaf and the universality of the $H^{1}$ -obstruction class. Read in the opposite direction, the Bishop-Remmert-Narasimhan embedding theorem identifies Stein manifolds with closed complex submanifolds of $C^{N}$ — this is the central insight that converts the abstract three-axiom definition into a concrete geometric realisation. The Levi-problem equivalence further identifies Stein-ness with pseudoconvexity and with domains-of-holomorphy on open subsets of $C^{n}$ , and the Oka principle promotes the dimension-one Cousin II solvability into a homotopy-theoretic identification of holomorphic and continuous bundle classifications on every Stein source in every complex dimension. Putting these together, the bridge is the Hörmander weighted $L^{2}$ -existence theorem for $\overset{ˉ}{\partial}$ : this is exactly the analytic engine that powers Theorem B in arbitrary complex dimension, generalises the dimension-one Behnke-Stein construction, and identifies the Stein property with the Levi-form positivity of the exhaustion. The dimension-one specialisation 06.09.02 is the cleanest expression because the syzygy collapses to length one and the Hartogs obstruction is absent; the higher-dimensional case in this survey builds toward the full Cartan-Serre-Grauert framework, with elliptic-h-principle modern refinements (Gromov 1989, Forstnerič 2017) as the contemporary frontier.

Full proof set [Master]

The survey is synoptic: full proofs of the named theorems live in either the dimension-one specialisations of 06.09.02–06.09.07 or in the canonical references Hörmander 1973 / Grauert-Remmert 1979 / Forstnerič 2017. The proofs below either reprove dimension-one statements at the higher-dimensional level of generality, or state the theorem and cite the canonical proof.

Proposition (the polydisc is Stein). The polydisc $Δ^{n} = {(z_{1}, \dots, z_{n}) : ∣ z_{i} ∣ < 1}$ is a Stein manifold for every $n \geq 1$ .

Proof. Define $φ : Δ^{n} \to R$ by $φ (z) = \sum_{i = 1}^{n} - lo g (1 - ∣ z_{i} ∣^{2})$ . The function is smooth on $Δ^{n}$ , goes to infinity as $z$ approaches the boundary $\partial Δ^{n} = ⋃_{i} {∣ z_{i} ∣ = 1}$ , and is therefore an exhaustion. The Levi form is $$ i \partial \bar\partial \varphi(z) = \sum_{i=1}^n \frac{i , dz_i \wedge d\bar z_i}{(1 - |z_i|^2)^2}, $$ which is a sum of positive multiples of the standard Kähler form on each factor, hence strictly positive at every point. The function $φ$ is therefore a strictly plurisubharmonic exhaustion. By the equivalent characterisation of Stein-ness (strictly plurisubharmonic exhaustion ⇔ Stein, due to Grauert 1958 / Hörmander 1965), $Δ^{n}$ is Stein. $□$

Proposition (Cousin II on a contractible Stein). Let $X$ be a contractible Stein manifold (e.g., $C^{n}$ , $Δ^{n}$ , any convex bounded domain in $C^{n}$ ). Every Cousin II datum on $X$ is unconditionally solvable.

Proof. By the exponential exact sequence and Cartan-Serre Theorem B (vanishing of $H^{1}$ and $H^{2}$ of $O_{X}$ ), the long exact sequence gives $Pic (X) = H^{1} (X, O_{X}^{*}) ≅ H^{2} (X, Z)$ . A contractible space has vanishing integer cohomology in every positive degree: $H^{q} (X, Z) = 0$ for $q \geq 1$ and in particular $H^{2} (X, Z) = 0$ . So $Pic (X) = 0$ on a contractible Stein, every line bundle is holomorphically isomorphic to $O_{X}$ , and every Cousin II cocycle is a coboundary $g_{ij} = h_{i} / h_{j}$ with $h_{i} \in O^{*} (U_{i})$ . The global $f = f_{i} / h_{i}$ on $U_{i}$ assembles to a non-zero meromorphic function on $X$ realising the prescribed divisor data. $□$

Proposition (Cousin II failure on a non-contractible Stein). On $X = \mathbb{C}^ \times \mathbb{C}^* \subset \mathbb{C}^2$, Cousin II is not unconditionally solvable.*

Proof. $X$ is Stein (it is an open subset of $C^{2}$ admitting the strictly plurisubharmonic exhaustion $φ (z, w) = ∣ lo g ∣ z ∣ ∣^{2} + ∣ lo g ∣ w ∣ ∣^{2} + ∣ z ∣^{2} + ∣ w ∣^{2}$ , suitably normalised). Topologically, $X ≃ (S^{1})^{2}$ , so $H^{2} (X, Z) ≅ Z$ generated by the cup product $[d θ_{1}] \cup [d θ_{2}]$ . By the exponential sequence and Theorem B, $Pic (X) ≅ H^{2} (X, Z) = Z$ . The non-zero generator is realised by an explicit divisor: the divisor ${z + w = 0} \cap X$ (the diagonal in $X$ via the multiplicative identification) defines a holomorphic line bundle with non-vanishing Chern class, equal to the generator. Cousin II data with this Chern class are not solvable — equivalent to the non-zero line bundle being isomorphic to $O_{X}$ , which it is not. $□$

Proposition (Bishop-Remmert embedding for $C^{n}$ ). The identity $id : C^{n} ↪ C^{n}$ is a proper holomorphic embedding. So the Bishop-Remmert dimension bound is $N = n$ for $X = C^{n}$ , which is sharper than the general bound $N = 2 n + 1$ .

Proof. The identity map is holomorphic, embedding, and proper (because $C^{n}$ is closed in itself). The minimum embedding dimension of a Stein manifold of complex dimension $n$ ranges between $n$ (achieved by $C^{n}$ ) and $2 n + 1$ (Remmert's sharp upper bound). The Eliashberg-Gromov refinement $⌊ 3 n /2 ⌋ + 1$ is achieved on certain Stein manifolds where Remmert's bound is loose. The point of the embedding theorem is the uniformity of $N$ across all Stein manifolds of given dimension, not the achievement of the minimum on each individual manifold. $□$

Theorem (Cartan-Serre Theorem B in arbitrary complex dimension, full statement). Statement and proof as in the Intermediate-tier Key theorem section.

Proof. See the Intermediate-tier proof (Steps 1–5), with the following higher-dimensional details added. Step 2's Hörmander $L^{2}$ -existence uses the weighted $L^{2} (ϕ)$ -space on $X_{n}$ with weight $ϕ = K φ$ for $K ≫ 0$ chosen so that the Levi form $i \partial \overset{ˉ}{\partial} ϕ$ dominates a fixed positive $(1, 1)$ -form on $X_{n}$ ; the existence proof uses the Bochner-Kodaira-Nakano identity to control the closed-range property of $\overset{ˉ}{\partial}$ . Step 3's syzygy induction uses Hilbert's theorem on regular local rings: the local ring $O_{X, x}$ at a point $x$ of complex dimension $n$ is regular noetherian of Krull dimension $n$ , hence has global homological dimension $n$ , so every coherent sheaf admits a free resolution of length at most $n$ . Step 4's limit passage uses Cartan's Runge-type approximation on Stein subdomains: on a relatively compact Stein open $X_{n} ⋐ X_{n + 1}$ in a larger Stein $X_{n + 1}$ , the restriction $Γ (X_{n + 1}, F) \to Γ (X_{n}, F)$ has dense image in the Fréchet topology. This is the higher-dimensional Runge approximation, proved via the same $\overset{ˉ}{\partial}$ -cutoff argument as the dimension-one case in 06.09.07. The Mittag-Leffler condition on the inverse system ${Γ (X_{n}, F)}$ then commutes the inverse limit with cohomology. $□$

Corollary (Theorem A from Theorem B). Under the hypotheses of Theorem B, the natural evaluation map $Γ (X, F) \otimes_{Γ (X, O_{X})} O_{X, x} \to F_{x}$ is surjective for every $x \in X$ .

Proof. Given $x \in X$ and $σ \in F_{x}$ , extend $σ$ to a section $\tilde{σ} \in Γ (U, F)$ on a neighbourhood $U$ of $x$ . Define a sheaf morphism $O_{X} \to F$ by $1 \mapsto \tilde{σ}$ on $U$ and zero outside (more precisely: take $G = O_{X}$ with a coherent quotient $O_{X} \to F$ realising $\tilde{σ}$ on $U$ , formed as the cokernel of the annihilator-sheaf inclusion). The kernel $K$ of this morphism is coherent. The short exact sequence $0 \to K \to O_{X} \to im \to 0$ gives a long exact sequence in cohomology $$ \Gamma(X, \mathcal{O}_X) \to \Gamma(X, \mathrm{im}) \to H^1(X, \mathcal{K}) \to \cdots. $$ By Theorem B applied to the coherent $K$ , $H^{1} (X, K) = 0$ , so the connecting map vanishes and $Γ (X, O_{X}) \to Γ (X, im)$ is surjective. Specialising at $x$ , the global-section evaluation surjects onto a generator $\tilde{σ}$ of $F_{x}$ , giving Theorem A. $□$

Theorem (Oka principle, Grauert form). Let $X$ be a Stein manifold and $G$ a complex Lie group. The classification maps ${holomorphic principal G -bundles on X} / ≅\to {continuous principal G -bundles on X} / ≅$ is a bijection.

Proof. Stated without proof here; full proof in Grauert 1958 Math. Ann. 135, with simplifications in Grauert-Remmert 1979 Theory of Stein Spaces Ch. V and modern reformulations in Forstnerič 2017 Stein Manifolds and Holomorphic Mappings Ch. 5. The argument: surjectivity uses the existence of holomorphic gauge transformations connecting any continuous local frame to a holomorphic one — proved by an induction on a Stein exhaustion with Cartan-Serre Theorem B applied to the gauge-transition sheaf at each step. Injectivity uses the deformation argument: any continuous gauge transformation between two holomorphic bundles is homotopic to a holomorphic one, with the homotopy constructed by a $\overset{ˉ}{\partial}$ -solvability argument applied to the family of continuous deformations. $□$

Connections [Master]

Cartan's Theorems A and B for Stein Riemann surfaces 06.09.02. The dimension-one specialisation of the Cartan-Serre theory surveyed here. The dimension-one case in 06.09.02 uses a syzygy of length one (the local rings $O_{X, x}$ on a Riemann surface are regular of Krull dimension one), with Hörmander $L^{2}$ -existence on a strictly subharmonic exhaustion as the analytic engine. The higher-dimensional case in this survey requires a syzygy of length $n = dim_{C} X$ and the Hörmander weighted $L^{2}$ -existence theorem in arbitrary complex dimension. The structural shape — exhaust by Stein subdomains, solve $\overset{ˉ}{\partial}$ on each, limit-pass with Mittag-Leffler / Cartan-Runge approximation — is identical in every dimension.
Behnke-Stein theorem 06.09.03. The dimension-one ancestor of the Stein theory: every non-compact Riemann surface is Stein. In dimension one, the Stein property is the only condition compatible with non-compactness, so the abstract Stein-manifold definition collapses to non-compact. In dimension $n \geq 2$ , the Stein property is a real restriction: most open subsets of $C^{n}$ are Stein, but the Hartogs shells and other non-pseudoconvex domains are not. Behnke-Stein 1949 (Math. Ann. 120, 430–461) was the analytic foundation of the entire theory, predating the abstract Stein-manifold definition by two years.
Cousin I (additive) 06.09.04. The dimension-one specialisation of the Cousin I problem on a Stein manifold. The cohomological obstruction $[f_{i} - f_{j}] \in H^{1} (X, O_{X})$ vanishes by Theorem B on every Stein manifold in every complex dimension. In dimension one, Behnke-Stein automatically gives Stein-ness for non-compact RS; in higher dimension, the Stein hypothesis is part of the data and must be verified (e.g., via a strictly plurisubharmonic exhaustion, the Levi-problem equivalence, or the Bishop-Remmert embedding).
Cousin II (multiplicative) 06.09.05. The dimension-one specialisation of the Cousin II problem. The multiplicative cocycle obstruction lives in $H^{1} (X, O_{X}^{*}) = Pic (X) ≅ H^{2} (X, Z)$ on a Stein manifold. In dimension one (non-compact RS), $H^{2} (X, Z) = 0$ for topological reasons, so Cousin II is unconditionally solvable. In higher dimension, $H^{2} (X, Z)$ can be non-zero on Stein manifolds (e.g., $C^{*} \times C^{*} \subset C^{2}$ ), and Cousin II is conditionally solvable, with the Chern class as the obstruction.
Mittag-Leffler on RS 06.09.06. The dimension-one specialisation of the Mittag-Leffler problem. Prescribed Laurent tails at a discrete set, realised by a global meromorphic function. On every Stein manifold in every dimension, the Mittag-Leffler obstruction $δ : Γ (X, P_{X}) \to H^{1} (X, O_{X}) = 0$ vanishes by Theorem B. The dimension-one statement on a non-compact RS is unconditionally solvable by Behnke-Stein; the higher-dimensional statement on a Stein manifold of any complex dimension is solvable by Cartan-Serre.
Runge approximation on RS 06.09.07. The dimension-one specialisation of the Oka-Weil approximation theorem. In dimension one, Runge approximation on a non-compact RS is the surjectivity-on-restriction theorem for holomorphic functions, equivalent to the compactly-supported $H^{1}$ -vanishing dual to Theorem B. In higher dimension, the Oka-Weil theorem on polynomially convex compact subsets of $C^{n}$ generalises the planar Runge theorem; the abstract Stein-manifold version (Hörmander 1965, Several Complex Variables) holds on every $O (X)$ -convex compact subset of a Stein manifold $X$ .
Holomorphic functions of several complex variables 06.07.01. The framework that makes higher-dimensional Stein theory possible. Without the multivariate machinery — Cauchy formulas in several variables, the Hartogs extension theorem, the local $\overset{ˉ}{\partial}$ -Poincaré lemma in $n$ variables — the abstract Stein-manifold definition would be empty. The unit 06.07.01 supplies the foundational tools that the survey here builds on.
Hartogs phenomenon 06.07.02. The obstruction to Stein-ness in dimension $\geq 2$ . On a Hartogs shell or any non-pseudoconvex open subset of $C^{n}$ , every holomorphic function extends to a strictly larger open, breaking holomorphic convexity. The Hartogs phenomenon is the geometric content of the failure of the Levi-problem equivalence on non-pseudoconvex domains. The dimension-one analogue (annuli, punctured discs) is absent: in dimension one, every open set is a domain of holomorphy, and Behnke-Stein gives Stein-ness automatically on non-compact RS.
Sheaf cohomology survey on Riemann surfaces 06.04.07. The cohomological framework — four pictures of sheaf cohomology, comparison theorems, GAGA — on which Cartan-Serre Theorem B lives. The survey 06.04.07 gives the compact-Kähler / Riemann-surface side; the present survey gives the Stein / non-compact side; together they cover the two extremal regimes of complex analytic geometry.
Riemann-Roch theorem for compact Riemann surfaces 06.04.01. The compact-curve dimension-counting theorem, dual to the Stein-manifold cohomological-vanishing theorem. On a compact Riemann surface, Riemann-Roch gives a sharp two-sided dimension count; on a Stein manifold, Cartan-Serre gives an unconditional vanishing. The two regimes — compact-Kähler with finite-dimensional cohomology, and non-compact-Stein with vanishing cohomology — are the structural extremes of complex analytic geometry, and every intermediate complex manifold (non-compact non-Stein, compact non-Kähler, partially convex) sits between them.
Symplectic / Weinstein structures (pending future unit on Cieliebak-Eliashberg). Every Stein manifold carries a canonical Weinstein structure compatible with the cotangent symplectic form: the strictly plurisubharmonic exhaustion $φ$ gives the Liouville one-form $- d^{c} φ$ , the gradient of $φ$ gives the Liouville vector field, and the level sets are contact. This is the bridge from complex analysis to symplectic topology, formalised in Cieliebak-Eliashberg 2012 From Stein to Weinstein and Back. Stein manifolds appear as the source side of many h-principles in symplectic topology, including the Forstnerič-Gromov elliptic h-principle on holomorphic maps Stein → elliptic target.

Historical & philosophical context [Master]

Karl Stein defined the abstract manifold concept that bears his name in 1951 in Analytische Funktionen mehrerer komplexer Veränderlichen zu vorgegebenen Periodizitätsmoduln und das zweite Cousinsche Problem ^{[Stein 1951]} (Math. Ann. 123, 201–222). Stein had been working on the period-prescription problem in $C^{n}$ posed by Cousin in 1895: given prescribed periodicity data on a covering of $C^{n}$ minus a discrete set, find a global meromorphic function realising the data. Stein isolated the three axioms — holomorphic convexity, holomorphic separability, holomorphic spread — as the abstract geometric conditions sufficient for the period problem to be solvable for every coherent-sheaf datum. The motivation was simultaneously the dimension-one Behnke-Stein theorem of 1949 ^{[Behnke-Stein 1949]} and the dimension-two Oka 1939 Cousin II solvability.

Henri Cartan's 1951–53 séminaire at the École Normale Supérieure ^{[Cartan 1951–53]} and the Cartan-Serre 1953 Comptes Rendus note Un théorème de finitude concernant les variétés analytiques compactes ^{[Cartan-Serre 1953]} (C. R. Acad. Sci. Paris 237, 128–130) extracted the cohomological content: Theorem A (global sections generate stalks) and Theorem B (higher cohomology vanishes) for every coherent analytic sheaf on a Stein manifold in arbitrary complex dimension. The séminaire was a watershed: it absorbed Stein's three axioms, the Behnke-Stein dimension-one theorem, Oka's coherence theorem (Oka 1950), and the Cousin problem solvability into a single cohomological identity, valid in every complex dimension. Cartan and Serre's Faisceaux Algébriques Cohérents paper of 1955 (Serre's habilitation; Annals of Math. 61) is the algebraic-geometric counterpart, with vanishing on affine schemes replacing vanishing on Stein manifolds — the GAGA bridge that Serre would establish in 1956.

Reinhold Remmert in 1956 ^{[Remmert 1956]} (C. R. Acad. Sci. Paris 243, 118–121) proved the embedding theorem: every Stein manifold of complex dimension $n$ embeds properly into $C^{2 n + 1}$ . Raghavan Narasimhan in 1960 ^{[Narasimhan 1960]} (Amer. J. Math. 82, 917–934) extended the result to Stein analytic spaces with singularities; Errett Bishop in 1961 ^{[Bishop 1961]} (Amer. J. Math. 83, 209–242) refined the construction. The combined result — known today as the Bishop-Remmert-Narasimhan embedding theorem — identifies Stein manifolds with closed complex submanifolds of $C^{N}$ for some $N$ , completing the geometric characterisation begun with Stein's three axioms in 1951. Yakov Eliashberg and Mikhael Gromov in 1992 ^{[Eliashberg-Gromov 1992]} (Ann. of Math. 136, 123–135) sharpened the embedding dimension to $N = ⌊ 3 n /2 ⌋ + 1$ for $n \geq 2$ , optimal in general.

Kiyoshi Oka's papers from 1936 to 1953 (collected in Sur les fonctions analytiques de plusieurs variables, I–VIII, J. Sci. Hiroshima Univ. Ser. A and Tôhoku Math. J.) established the analytic foundations of the entire theory: the Cousin II solvability on domains of holomorphy (Oka 1939) ^{[Oka 1939]}, the coherence of the structure sheaf (Oka 1950), and the Levi problem solution in dimension two (Oka 1942) ^{[Oka 1942]}. Heinz Bremermann (1954) and François Norguet (1954) independently solved the Levi problem in arbitrary complex dimension simultaneously with Oka 1953 ^{[Bremermann-Norguet-Oka]}, identifying pseudoconvex domains with Stein manifolds on open subsets of $C^{n}$ . Hans Grauert's 1958 paper Analytische Faserungen über holomorph-vollständigen Räumen ^{[Grauert 1958]} (Math. Ann. 135, 263–273) extended Oka's principle to principal $G$ -bundles for any complex Lie group $G$ , founding the Oka theory that Forstnerič and Larusson developed into modern h-principle theory.

Lars Hörmander's 1965 paper $L^{2}$ -estimates and existence theorems for the $\overset{ˉ}{\partial}$ operator ^{[Hörmander 1965]} (Acta Math. 113, 89–152) supplied the modern analytic engine: a self-contained PDE proof of Theorem B on any strictly plurisubharmonic exhaustion, using weighted $L^{2}$ -norms and the Bochner-Kodaira-Nakano identity. Hörmander's textbook An Introduction to Complex Analysis in Several Variables ^{[Hörmander HSCV]} (North-Holland 1973) gave the canonical textbook treatment of Stein theory, replacing the Cartan séminaire's exhaustion-by-Stein-subdomains with the PDE proof. Hans Grauert and Reinhold Remmert's Theory of Stein Spaces ^{[Grauert-Remmert 1979]} (Springer Grundlehren 236, 1979) is the comprehensive monograph, extending Stein theory to singular analytic spaces. Franc Forstnerič's Stein Manifolds and Holomorphic Mappings ^{[Forstnerič 2017]} (Springer Ergebnisse 56, 2nd ed. 2017) is the modern reference for the Oka principle and elliptic h-principles on Stein sources, building on Mikhael Gromov's 1989 Oka's principle for holomorphic sections of elliptic bundles ^{[Gromov 1989]} (J. Amer. Math. Soc. 2, 851–897).

Bibliography [Master]

@article{Stein1951,
  author  = {Stein, Karl},
  title   = {Analytische {Funktionen} mehrerer komplexer {Ver{\"a}nderlichen} zu vorgegebenen {Periodizit{\"a}tsmoduln} und das zweite {Cousinsche} {Problem}},
  journal = {Math. Ann.},
  volume  = {123},
  year    = {1951},
  pages   = {201--222}
}

@article{CartanSerre1953,
  author  = {Cartan, Henri and Serre, Jean-Pierre},
  title   = {Un th{\'e}or{\`e}me de finitude concernant les vari{\'e}t{\'e}s analytiques compactes},
  journal = {C. R. Acad. Sci. Paris},
  volume  = {237},
  year    = {1953},
  pages   = {128--130}
}

@misc{CartanSeminaire,
  author       = {Cartan, Henri},
  title        = {S{\'e}minaire {Henri Cartan}, {\'E}cole {Normale Sup{\'e}rieure}},
  howpublished = {Expos{\'e}s 1951--1953 on coherent analytic sheaves and Stein manifolds},
  year         = {1951--1953}
}

@article{Remmert1956,
  author  = {Remmert, Reinhold},
  title   = {Sur les espaces analytiques holomorphiquement s{\'e}parables et holomorphiquement convexes},
  journal = {C. R. Acad. Sci. Paris},
  volume  = {243},
  year    = {1956},
  pages   = {118--121}
}

@article{Narasimhan1960,
  author  = {Narasimhan, Raghavan},
  title   = {Imbedding of holomorphically complete complex spaces},
  journal = {Amer. J. Math.},
  volume  = {82},
  year    = {1960},
  pages   = {917--934}
}

@article{Bishop1961,
  author  = {Bishop, Errett},
  title   = {Mappings of partially analytic spaces},
  journal = {Amer. J. Math.},
  volume  = {83},
  year    = {1961},
  pages   = {209--242}
}

@article{EliashbergGromov1992,
  author  = {Eliashberg, Yakov and Gromov, Mikhael},
  title   = {Embeddings of {Stein} manifolds of dimension $n$ into the affine space of dimension $3n/2 + 1$},
  journal = {Ann. of Math. (2)},
  volume  = {136},
  year    = {1992},
  pages   = {123--135}
}

@article{BehnkeStein1949,
  author  = {Behnke, Heinrich and Stein, Karl},
  title   = {{Entwicklung analytischer Funktionen auf Riemannschen Fl{\"a}chen}},
  journal = {Math. Ann.},
  volume  = {120},
  year    = {1949},
  pages   = {430--461}
}

@article{Oka1939,
  author  = {Oka, Kiyoshi},
  title   = {Sur les fonctions analytiques de plusieurs variables {III}: {Deuxi{\`e}me} probl{\`e}me de {Cousin}},
  journal = {J. Sci. Hiroshima Univ. Ser. A},
  volume  = {9},
  year    = {1939},
  pages   = {7--19}
}

@article{Oka1942,
  author  = {Oka, Kiyoshi},
  title   = {Sur les fonctions analytiques de plusieurs variables {VI}: Domaines pseudoconvexes},
  journal = {T{\^o}hoku Math. J.},
  volume  = {49},
  year    = {1942},
  pages   = {15--52}
}

@article{Grauert1958,
  author  = {Grauert, Hans},
  title   = {{Analytische Faserungen {\"u}ber holomorph-vollst{\"a}ndigen R{\"a}umen}},
  journal = {Math. Ann.},
  volume  = {135},
  year    = {1958},
  pages   = {263--273}
}

@article{Hormander1965,
  author  = {H{\"o}rmander, Lars},
  title   = {$L^2$ estimates and existence theorems for the $\bar\partial$ operator},
  journal = {Acta Mathematica},
  volume  = {113},
  year    = {1965},
  pages   = {89--152}
}

@book{HormanderHSCV,
  author    = {H{\"o}rmander, Lars},
  title     = {An Introduction to Complex Analysis in Several Variables},
  edition   = {Third},
  publisher = {North-Holland},
  year      = {1990}
}

@book{GrauertRemmert1979,
  author    = {Grauert, Hans and Remmert, Reinhold},
  title     = {Theory of {Stein} Spaces},
  series    = {Grundlehren der mathematischen Wissenschaften},
  volume    = {236},
  publisher = {Springer},
  year      = {1979}
}

@article{Gromov1989,
  author  = {Gromov, Mikhael},
  title   = {{Oka's} principle for holomorphic sections of elliptic bundles},
  journal = {J. Amer. Math. Soc.},
  volume  = {2},
  year    = {1989},
  pages   = {851--897}
}

@book{Forstneric2017,
  author    = {Forstneri{\v{c}}, Franc},
  title     = {Stein Manifolds and Holomorphic Mappings: The Homotopy Principle in Complex Analysis},
  series    = {Ergebnisse der Mathematik und ihrer Grenzgebiete},
  volume    = {56},
  edition   = {Second},
  publisher = {Springer},
  year      = {2017}
}

@book{ForsterRiemannSurfaces,
  author    = {Forster, Otto},
  title     = {Lectures on {Riemann} Surfaces},
  series    = {Graduate Texts in Mathematics},
  volume    = {81},
  publisher = {Springer},
  year      = {1981}
}

@article{Hartogs1906,
  author  = {Hartogs, Friedrich},
  title   = {Zur {Theorie} der analytischen {Funktionen} mehrerer unabh{\"a}ngiger {Ver{\"a}nderlichen}},
  journal = {Math. Ann.},
  volume  = {62},
  year    = {1906},
  pages   = {1--88}
}

@article{Cousin1895,
  author  = {Cousin, Pierre},
  title   = {Sur les fonctions de $n$ variables complexes},
  journal = {Acta Math.},
  volume  = {19},
  year    = {1895},
  pages   = {1--62}
}

@book{CieliebakEliashberg2012,
  author    = {Cieliebak, Kai and Eliashberg, Yakov},
  title     = {From {Stein} to {Weinstein} and Back: Symplectic Geometry of Affine Complex Manifolds},
  series    = {American Mathematical Society Colloquium Publications},
  volume    = {59},
  publisher = {AMS},
  year      = {2012}
}

Prerequisites

06.09.02
06.09.03
06.09.04
06.09.05
06.09.06
06.09.07
06.07.01
06.07.02

Tier anchors

beginner: Hörmander *An Introduction to Complex Analysis in Several Variables* Ch. II informal; Forster *Lectures on Riemann Surfaces* §29 informal (dimension-one ancestor)
intermediate: Hörmander *An Introduction to Complex Analysis in Several Variables* Ch. IV–V; Grauert-Remmert *Theory of Stein Spaces* Ch. IV–V
master: Stein 1951 *Analytische Funktionen mehrerer komplexer Veränderlichen zu vorgegebenen Periodizitätsmoduln und das zweite Cousinsche Problem* (Math. Ann. 123, originator: Stein-manifold definition); Cartan 1951–53 séminaire (originator: Theorems A and B); Cartan-Serre 1953 *Un théorème de finitude concernant les variétés analytiques compactes* (CRAS 237, originator: Theorem B in arbitrary complex dimension); Remmert 1956 *Sur les espaces analytiques holomorphiquement séparables et holomorphiquement convexes* (CRAS 243, originator: Remmert embedding); Bishop 1961 *Mappings of partially analytic spaces* (Amer. J. Math. 83, originator: sharp embedding dimension); Narasimhan 1960 *Imbedding of holomorphically complete complex spaces* (Amer. J. Math. 82, originator: sharp embedding for Stein spaces); Behnke-Stein 1949 *Entwicklung analytischer Funktionen auf Riemannschen Flächen* (Math. Ann. 120, originator: dimension-one ancestor); Hörmander *An Introduction to Complex Analysis in Several Variables* Ch. IV–V (textbook); Grauert-Remmert *Theory of Stein Spaces* (Grundlehren 236, canonical reference); Forstnerič *Stein Manifolds and Holomorphic Mappings* (2nd ed., Springer 2017, Oka principle); Gromov 1989 *Oka's principle for holomorphic sections of elliptic bundles* (J. Amer. Math. Soc. 2)

References

TODO_REF
Stein 1951 — Analytische Funktionen mehrerer komplexer Veränderlichen zu vorgegebenen Periodizitätsmoduln und das zweite Cousinsche Problem · Math. Ann. 123, 201–222 (originator: the axiomatic definition of a Stein manifold via holomorphic convexity, separation, and local-coordinate spreading)
TODO_REF
Cartan 1951–53 séminaire — Théorèmes A et B · Séminaire Cartan, ENS, exposés 1951–53 (originator: Theorems A and B for Stein manifolds; coherent analytic sheaves)
TODO_REF
Cartan-Serre 1953 — Un théorème de finitude concernant les variétés analytiques compactes · C. R. Acad. Sci. Paris 237, 128–130 (originator: Theorem B in arbitrary complex dimension; Schwartz finiteness on compact analytic spaces)
TODO_REF
Remmert 1956 — Sur les espaces analytiques holomorphiquement séparables et holomorphiquement convexes · C. R. Acad. Sci. Paris 243, 118–121 (originator: Stein manifolds embed properly into $\mathbb{C}^N$; dimension bound $N = 2n + 1$)
TODO_REF
Narasimhan 1960 — Imbedding of holomorphically complete complex spaces · Amer. J. Math. 82, 917–934 (sharpened the Remmert embedding for Stein spaces with singularities)
TODO_REF
Bishop 1961 — Mappings of partially analytic spaces · Amer. J. Math. 83, 209–242 (further refinement of the embedding dimension; the present-day form of the Bishop-Remmert-Narasimhan embedding theorem)
TODO_REF
Eliashberg-Gromov 1992 — Embeddings of Stein manifolds · Ann. of Math. 136, 123–135 (sharp dimension $N = \lfloor 3n/2 \rfloor + 1$ for the Stein embedding)
TODO_REF
Behnke-Stein 1949 — Entwicklung analytischer Funktionen auf Riemannschen Flächen · Math. Ann. 120, 430–461 (dimension-one ancestor: every non-compact Riemann surface is Stein)
TODO_REF
Oka 1939 — Sur les fonctions analytiques de plusieurs variables III: Deuxième problème de Cousin · J. Sci. Hiroshima Univ. Ser. A 9 (originator: Oka's principle on the equivalence of continuous and holomorphic line-bundle classifications on Stein manifolds)
TODO_REF
Grauert 1958 — Analytische Faserungen über holomorph-vollständigen Räumen · Math. Ann. 135, 263–273 (originator: Oka principle for principal holomorphic bundles with complex Lie group fibre on a Stein space)
TODO_REF
Gromov 1989 — Oka's principle for holomorphic sections of elliptic bundles · J. Amer. Math. Soc. 2, 851–897 (Gromov's h-principle reformulation; modern Oka theory on elliptic targets)
TODO_REF
Hörmander — An Introduction to Complex Analysis in Several Variables · North-Holland 1973, Ch. IV–V (textbook treatment of Stein manifolds via the $L^2$-method for $\bar\partial$)
TODO_REF
Hörmander 1965 — $L^2$-estimates and existence theorems for the $\bar\partial$-operator · Acta Math. 113, 89–152 (modern $L^2$-PDE proof of Theorem B on Stein manifolds)
TODO_REF
Grauert-Remmert — Theory of Stein Spaces · Springer Grundlehren 236, 1979 (canonical reference for Stein theory and coherent analytic sheaves)
TODO_REF
Forstnerič — Stein Manifolds and Holomorphic Mappings · Springer Ergebnisse 56, 2nd ed. 2017 (modern Oka theory and elliptic h-principles on Stein source)
TODO_REF
Hartogs 1906 — Zur Theorie der analytischen Funktionen mehrerer unabhängiger Veränderlichen · Math. Ann. 62, 1–88 (originator of the Hartogs extension phenomenon, the obstruction to Stein-ness on shells in $\mathbb{C}^n$ for $n \geq 2$)
TODO_REF
Oka 1942 — Sur les fonctions analytiques de plusieurs variables VI: Domaines pseudoconvexes · Tôhoku Math. J. 49, 15–52 (originator: Levi problem; solved in dimension two)
TODO_REF
Bremermann 1954 / Norguet 1954 / Oka 1953 — Levi problem in arbitrary dimension · (three independent simultaneous proofs that pseudoconvex domains in $\mathbb{C}^n$ are domains of holomorphy, hence Stein)

Reviewer

TBD

Estimated time

beginner: 14m
intermediate: 38m
master: 70m