Loading field map
Complex geometry is the study of geometric structures modeled on the complex numbers. Its primary objects are complex manifolds: spaces that locally resemble the familiar complex coordinate space ℂⁿ, with the requirement that the "patches" used to assemble the global space are glued together by holomorphic (complex-differentiable) maps. Because holomorphic functions are far more rigid than smooth ones, complex manifolds carry a much richer structure than ordinary differentiable manifolds, and the field is largely devoted to understanding how that rigidity shapes global geometry.
A complex manifold is defined by an atlas of coordinate charts whose transition functions are holomorphic. This single condition has profound consequences. On a compact complex manifold, every global holomorphic function is constant—there is no nonconstant complex-analytic analogue of a smooth bump function. The geometry must therefore be studied through more subtle invariants: holomorphic line bundles, cohomology groups, and the curvature of Hermitian metrics.
The central tension in the field is between the local flexibility of complex analysis and the global rigidity imposed by compactness. Locally, holomorphic maps are extremely constrained (they are power series), yet globally, compact complex manifolds can be assembled in surprisingly many ways. Classifying these manifolds, and understanding which geometric structures they admit, is the field's enduring project.
A second fundamental object is the complex projective space ℂℙⁿ, the set of complex lines through the origin in ℂⁿ⁺¹. Projective algebraic varieties—the zero sets of homogeneous polynomial equations in ℂℙⁿ—form the most tractable and best-understood class of complex manifolds. A central question, the Kodaira embedding theorem, characterizes exactly which compact complex manifolds arise this way: those admitting a certain positive curvature form, called a Hodge metric. This theorem links the analytic world of Hermitian geometry to the algebraic world of polynomial equations, and much of complex geometry is organized around this bridge.
The analytic approach to complex geometry begins with the observation that a complex manifold carries a natural integrable almost-complex structure—a linear map J on each tangent space whose square is −1, encoding multiplication by i. A Hermitian metric is a Riemannian metric compatible with J, and its associated 2-form ω (the Kähler form) measures how the metric interacts with the complex structure.
A Kähler manifold is a complex manifold with a Hermitian metric whose Kähler form is closed (dω = 0). This condition is remarkably powerful. It implies that the metric is, to first order, indistinguishable from the flat Euclidean metric on ℂⁿ, and it forces strong topological restrictions: the odd-dimensional Betti numbers are even, and the Hodge decomposition splits the de Rham cohomology into holomorphic pieces. Projective varieties, when equipped with the metric induced from the Fubini–Study metric on ℂℙⁿ, are always Kähler.
The analytic tradition reached a watershed with the Calabi conjecture, resolved by Shing-Tung Yau. The conjecture asked whether a given topological class (a Kähler class) could be represented by a Kähler metric with prescribed Ricci curvature. Yau's solution showed that, under a necessary topological condition, the answer is yes. This result turned complex geometry into a powerful tool for constructing metrics with controlled curvature, and it had immediate consequences in algebraic geometry, including the proof of the Miyaoka–Yau inequality for surfaces and the existence of Kähler–Einstein metrics on many manifolds.
The analytic tradition's methods are fundamentally nonlinear partial differential equations. The central technical tool is the continuity method: one deforms a known solution to a desired equation through a family of equations, using a priori estimates to show the deformation can proceed. This approach reached its culmination in the work on the Kähler–Ricci flow, a parabolic equation that evolves a Kähler metric toward a canonical representative of its cohomology class. The flow, developed by Gang Tian and others, provides a dynamical way to understand the existence and uniqueness of canonical metrics.
The algebraic tradition approaches complex manifolds through their function fields and cohomology. Its foundational result is the Kodaira classification of compact complex surfaces, completed by Kunihiko Kodaira in the 1960s. The classification organizes surfaces by their Kodaira dimension, a birational invariant measuring how many independent meromorphic functions the surface admits. The classification divides surfaces into four broad classes: rational and ruled surfaces (which are birationally simple), elliptic surfaces (which fiber over a curve with elliptic fibers), surfaces of general type (which are "maximally" complex), and the special case of K3 surfaces (which have trivial canonical bundle).
This classification is not merely a list; it is a structural theorem. It shows that every compact complex surface can be obtained from one of a small number of building blocks through blow-ups (replacing a point by a projective line) and fibrations. The classification also reveals a striking dichotomy: most surfaces are either "ruled" (covered by rational curves) or "of general type" (whose canonical bundle is ample). The borderline cases—elliptic surfaces and K3 surfaces—are precisely where the most interesting geometry lives.
The algebraic tradition also developed the theory of moduli spaces: parameter spaces for complex structures on a fixed topological manifold. The key insight, due to Kodaira and Donald Spencer, is that infinitesimal deformations of a complex structure are classified by the first cohomology group H¹(X, $T_X$) with values in the holomorphic tangent bundle. When this group vanishes, the deformation theory is unobstructed and the moduli space is smooth. When it does not vanish, obstructions can appear, and the moduli space can be singular or have multiple branches. The study of moduli spaces connects complex geometry to algebraic geometry through the construction of coarse moduli schemes, and it remains an active area, particularly for surfaces of general type and for higher-dimensional varieties.
Hodge theory is the meeting point of the analytic and algebraic traditions. It begins with the observation that on a compact Kähler manifold, the de Rham cohomology groups admit a decomposition into (p,q)-types, reflecting the complex structure. This decomposition is not merely topological; it varies holomorphically as the complex structure varies, giving rise to the period map, which sends a complex manifold to its Hodge structure.
The deepest open problem in the field, the Hodge conjecture, asks whether certain rational cohomology classes (those of type (p,p)) can always be represented by algebraic cycles—formal combinations of subvarieties. The conjecture remains unproved in general, and it is one of the Clay Millennium Problems. Its difficulty reflects a fundamental gap in our understanding: we know that algebraic cycles produce Hodge classes, but we do not know whether every Hodge class has a geometric origin.
A related but distinct problem is the study of variations of Hodge structure. When a family of complex manifolds is parameterized by a base space, the Hodge decomposition varies, and Griffiths transversality describes the infinitesimal constraints on this variation. This theory has been used to prove rigidity theorems—showing that certain manifolds cannot be deformed—and to construct period domains, which are homogeneous spaces for real Lie groups that parameterize abstract Hodge structures.
The classification of complex surfaces, completed in the 1960s, has resisted full generalization to higher dimensions. The minimal model program, developed primarily by Shigefumi Mori and his collaborators, provides the modern framework. The program seeks to replace a given projective variety by a "minimal model" whose canonical bundle is nef (numerically effective), or by a Fano fibration, through a sequence of birational operations called divisorial contractions and flips.
The minimal model program is fundamentally algebraic, but it has deep analytic consequences. The abundance conjecture, a central open problem, asserts that if the canonical bundle is nef, then some positive multiple is base-point-free, meaning it defines a holomorphic map. This conjecture is known in dimension three but open in higher dimensions. The program has been remarkably successful: it is known that minimal models exist for varieties of general type in all dimensions, and the termination of flips is known in dimension three and in many higher-dimensional cases.
The analytic counterpart to the minimal model program is the study of Kähler–Einstein metrics. A Kähler–Einstein metric is one whose Ricci form is proportional to the Kähler form. The existence of such a metric is governed by the sign of the first Chern class: positive (Fano), zero (Calabi–Yau), or negative (general type). The negative case was settled by Yau's solution of the Calabi conjecture. The zero case is also settled, yielding Ricci-flat metrics on Calabi–Yau manifolds. The positive case, for Fano manifolds, is more subtle: the existence of a Kähler–Einstein metric is equivalent to an algebro-geometric stability condition called K-stability, as conjectured by Yau and proved by Chen–Donaldson–Sun and, independently, by Tian. This result is a striking instance of the unity of the field: an analytic existence problem is solved by a purely algebraic criterion.
Contemporary complex geometry is characterized by the productive interaction of its analytic and algebraic strands. The Langlands program has entered the field through the geometric Langlands correspondence, which relates sheaves on moduli spaces of bundles to local systems, and through the study of character varieties. The Strominger–Yau–Zaslow conjecture, arising from mirror symmetry, proposes that Calabi–Yau manifolds admit special Lagrangian torus fibrations, a geometric structure that would explain the duality between symplectic and complex geometry. This conjecture remains open in general, but it has motivated extensive work on special Lagrangian submanifolds and on the structure of the moduli space of Calabi–Yau manifolds.
The field also continues to develop its technical foundations. The theory of singular complex spaces—spaces that are locally the zero sets of holomorphic functions but may have singularities—has become essential, both because moduli spaces are often singular and because the minimal model program produces singular varieties. The study of Hermitian metrics on singular spaces, and the extension of Hodge theory to this setting, is an active area.
A major open direction is the classification of Fano manifolds. While the minimal model program provides a framework, the classification of Fano manifolds in dimension four and higher is incomplete. The recent proof of the Kähler–Einstein existence criterion has sharpened the question: one must understand which Fano manifolds are K-stable, and this has led to deep connections with the theory of stability in geometric invariant theory.
Finally, the relationship between complex geometry and symplectic geometry remains a source of ongoing work. The two structures are compatible on a Kähler manifold, but not every symplectic manifold admits a compatible complex structure, and not every complex manifold admits a symplectic form. The question of which manifolds admit both—and how the two structures interact—is central to mirror symmetry and to the broader program of understanding the geometric structures that can coexist on a smooth manifold.
Complex geometry thus stands at the intersection of analysis, algebra, and topology. Its methods range from the most concrete (solving nonlinear PDEs) to the most abstract (derived categories and stability conditions), and its results connect to number theory, mathematical physics, and algebraic geometry. The field's central questions—what complex manifolds exist, how they deform, and which canonical metrics they admit—remain open in their most general forms, but the framework for addressing them is now remarkably coherent.