Loading field map
Algebraic topology is the branch of mathematics that studies topological spaces by assigning algebraic objects—such as groups, rings, or modules—to them in a way that respects continuous deformation. Its central insight is that certain qualitative features of a space, like the presence of holes or the way loops can be drawn on it, can be captured by algebraic structures that are easier to compute and compare than the spaces themselves. The field asks: when are two spaces essentially the same from the viewpoint of topology, and what algebraic invariants can tell them apart?
Topology concerns itself with properties of spaces that are preserved under homeomorphisms—continuous maps with continuous inverses. Two spaces are homeomorphic if one can be stretched, bent, or reshaped into the other without cutting or gluing. A coffee cup and a doughnut are homeomorphic because both have exactly one hole; the hole is a topological property, while the handle's shape is not.
Algebraic topology refines this idea by constructing invariants: algebraic objects assigned to each topological space such that homeomorphic spaces receive isomorphic algebraic objects. If two spaces have different invariants, they cannot be homeomorphic. The challenge is to make these invariants both powerful enough to distinguish many spaces and computable enough to be useful in practice.
The most fundamental invariants are homotopy groups and homology groups. Both detect holes, but in different dimensions and with different sensitivities. The first homotopy group, also called the fundamental group, captures information about loops: it records how loops in a space can be continuously deformed into one another, with the group operation given by concatenating loops. The fundamental group of a circle is the infinite cyclic group, reflecting the fact that a loop can wind around the circle any integer number of times. Higher homotopy groups generalize this to spheres of higher dimension, but they are notoriously difficult to compute.
Homology groups take a different approach. Instead of tracking deformations of maps from spheres, they build algebraic chains from the space's cells or simplices and measure the "holes" as cycles that are not boundaries. Homology is generally easier to compute than homotopy and has a richer algebraic structure, including a ring structure when one considers cohomology. The relationship between homotopy and homology—captured by the Hurewicz theorem, which states that the first nonzero homotopy group of a simply connected space maps isomorphically onto the corresponding homology group—is one of the field's foundational results.
The modern form of algebraic topology emerged in the mid-twentieth century when mathematicians realized that the assignment of groups to spaces is not merely a collection of isolated constructions but a functor: a systematic translation that carries not only spaces to groups but also continuous maps between spaces to homomorphisms between groups. This functorial perspective, codified in the language of category theory, transformed the field. It meant that topological questions could be translated into algebraic ones, and that the composition of continuous maps corresponds to the composition of group homomorphisms.
This shift allowed algebraic topologists to ask not just "what are the invariants of a space?" but "how do invariants behave under maps between spaces?" The answer led to powerful computational tools, such as the Mayer–Vietoris sequence, which computes the homology of a space from the homology of two overlapping subspaces, and the long exact sequence of a pair, which relates the homology of a space to that of a subspace. These tools made homology a practical instrument for calculation rather than a purely theoretical construct.
A central theme in algebraic topology is the classification of spaces up to homotopy equivalence, a weaker relation than homeomorphism. Two spaces are homotopy equivalent if there are continuous maps back and forth whose compositions are homotopic to the identity maps. Homotopy equivalence preserves all homotopy and homology groups, so algebraic invariants cannot distinguish homotopy equivalent spaces. For many purposes, homotopy equivalence is the right notion of "sameness" because it ignores differences that do not affect the algebraic invariants.
The classification problem asks: can we list all spaces up to homotopy equivalence, perhaps with some additional structure? For CW complexes—spaces built by attaching cells of increasing dimension—the answer is partially yes, thanks to Whitehead's theorem: a map between CW complexes that induces isomorphisms on all homotopy groups is a homotopy equivalence. This theorem reduces the classification problem to computing homotopy groups, which is far from solved. The homotopy groups of spheres, for instance, are known only partially and exhibit intricate patterns that remain an active area of research.
A major organizing insight in algebraic topology is the distinction between unstable and stable phenomena. As one suspends a space (roughly, forming a higher-dimensional analogue by adding a cone over it), its homotopy groups eventually settle into a pattern that no longer changes. The Freudenthal suspension theorem guarantees that, for sufficiently large dimensions, the homotopy groups stabilize. The stable homotopy groups of spheres form a graded abelian group that is a central object of study, and their computation is one of the field's great unsolved problems.
This stability led to the development of stable homotopy theory, which studies phenomena that are invariant under suspension. In this setting, one works with spectra—sequences of spaces connected by suspension maps—which serve as the stable analogue of spaces. Spectra support a rich algebraic structure, including a smash product that makes them into a symmetric monoidal category. This framework allows algebraic topologists to import techniques from algebra and homological algebra into topology, and it underlies many modern developments.
While homology groups are abelian groups, cohomology groups carry additional multiplicative structure: the cup product, which combines cohomology classes of a space to produce new ones. This product makes the cohomology of a space into a graded ring, a much finer invariant than the underlying groups. The cohomology ring of a space encodes information about how its holes interact with one another, and it can distinguish spaces that have identical homology groups.
The cup product arises naturally from the diagonal map of a space, and its existence reflects the fact that cohomology is contravariant: it reverses the direction of maps. This contravariance is essential for many applications, including the Poincaré duality theorem for closed manifolds, which relates the cohomology of a manifold in complementary dimensions. Poincaré duality is a cornerstone of the topology of manifolds and has deep connections to geometry and physics.
Many of the most powerful results in algebraic topology are obtained through spectral sequences: systematic procedures for computing complicated algebraic invariants from simpler ones through a sequence of successive approximations. A spectral sequence starts with an initial page of algebraic data and iteratively computes differentials that lead to the next page, eventually converging to the desired invariant. Spectral sequences are not invariants themselves but computational tools that organize information.
The most famous examples include the Serre spectral sequence, which computes the homology of a fibration from the homology of its base and fiber, and the Adams spectral sequence, which computes stable homotopy groups from cohomology operations. These tools are indispensable in practice, but they require care: spectral sequences often involve extension problems at the final stage, where the answer is determined only up to a group extension that must be resolved by additional arguments.
Algebraic topology has always maintained a close relationship with category theory, and this relationship deepened as the field matured. Many invariants can be understood as derived functors: algebraic constructions that measure the failure of exactness in a sequence of modules or groups. For example, the homology of a space can be computed as the derived functor of the tensor product, and cohomology as the derived functor of Hom. This perspective connects topology to homological algebra and allows techniques from one field to be imported into the other.
The categorical viewpoint also led to the development of model categories, which provide an abstract framework for doing homotopy theory in any category with suitable structure. Model categories formalize the notions of weak equivalence, fibration, and cofibration, and they allow algebraic topologists to define homotopy theories for algebraic objects, such as chain complexes or spectra, in a way that parallels the homotopy theory of spaces. This abstraction has been enormously influential, extending the reach of algebraic topology into algebra, geometry, and even mathematical physics.
The present landscape of algebraic topology is diverse, with several active research programs that build on the classical foundations while pushing into new territory.
Chromatic homotopy theory organizes stable homotopy groups according to a hierarchy of "heights," analogous to the chromatic filtration in algebraic geometry. This approach, inspired by the work of Jack Morava and developed by many others, has revealed deep connections between stable homotopy theory and formal group laws, leading to the nilpotence and periodicity theorems of Ethan Devinatz, Michael Hopkins, and Jeffrey Smith. These theorems show that certain families of elements in the stable homotopy groups of spheres are governed by periodic patterns, and they have transformed the study of stable homotopy.
Equivariant algebraic topology studies spaces equipped with group actions, seeking invariants that respect the symmetry. The Borel construction and equivariant cohomology are classical tools, but modern developments have introduced more refined invariants, such as genuine equivariant spectra, which carry information about all subgroups of the acting group. This area has found applications in representation theory and in the study of moduli spaces.
Derived algebraic geometry and topological modular forms represent a synthesis of algebraic topology with algebraic geometry. Topological modular forms are a spectrum whose homotopy groups encode information about elliptic curves and modular forms, and their construction required substantial new technology. This area exemplifies the modern trend of using spectra as a universal language that can encode sophisticated geometric and arithmetic information.
Open problems remain abundant. The computation of the stable homotopy groups of spheres is far from complete, and even the unstable homotopy groups of spheres are known only in low ranges. The Kervaire invariant problem, which asked whether certain high-dimensional manifolds exist, was resolved in all but one dimension by a combination of techniques from homotopy theory and surgery theory. The relationship between algebraic topology and mathematical physics, particularly through topological quantum field theories and string theory, continues to generate new questions and conjectures.
Despite its many sub-specialties, algebraic topology is unified by a single strategy: translate topological problems into algebraic ones, solve the algebraic problems, and translate the answers back. The field's power lies in the richness of the algebraic structures it employs—groups, rings, modules, spectra—and in the functoriality that makes these structures respond predictably to continuous maps. This strategy has proven remarkably robust, yielding invariants that are both computable and meaningful, and it continues to guide the field's development as it encounters new mathematical structures and new applications.