Manifolds 1 | Introduction and Topology [dark version]

TL;DR

The course motivates topology by the need to extend calculus from open regions to curved surfaces such as the sphere S², where constraints restrict motion to a surface.

Briefing Cornell Notes

Briefing

The course lays out a roadmap from topology to differentiable manifolds so calculus can be extended from flat domains to curved surfaces—an essential step for problems like optimization and motion constrained to a sphere. The central motivation is practical: when forces or constraints keep motion on a surface (for example, the 2-sphere S²), standard calculus tools for open regions no longer apply directly. The goal becomes building the right mathematical framework—starting with topology—to define what “open,” “neighborhood,” and later “differentiable” mean on spaces that may not live inside ordinary Euclidean space.

Part one begins with the sphere S² as the motivating example: it is the boundary of a ball in three dimensions and serves as a canonical 2-dimensional surface. From there, the course emphasizes that the key challenge is replacing metric-dependent calculus with a more flexible notion of structure. In metric spaces, open sets are defined using ε-balls (neighborhoods determined by a distance function d). But topology abstracts away the need for an explicit distance measurement. Instead of asking how far points are from each other, it asks which subsets should count as “open,” based on a set of rules.

That abstraction leads to the definition of a topology on a set X: it is a collection T of subsets of X (elements of the power set P(X)) that satisfies three axioms. First, both the empty set and the whole space X must be open. Second, the intersection of any two open sets must be open. Third, the union of an arbitrary collection of open sets—indexed by any index set I—must be open, meaning even infinitely many open sets can be combined while preserving openness.

With these axioms, “open” becomes a property defined relative to the chosen topology, not something inherited from a distance function. The course then illustrates two extreme examples. The indiscrete topology takes T to be {∅, X}, making only the trivial subsets open; it satisfies the axioms but offers little flexibility. The discrete topology takes T to be the entire power set P(X), declaring every subset open; it also satisfies the axioms but represents the opposite edge case. These examples are framed as boundary conditions that help clarify what the topology axioms actually permit.

The segment closes by signaling the next step: once the notion of open sets is in place, the course can proceed to use open sets as the foundation for further structure on spaces, eventually reaching differentiable manifolds, differential forms, and a generalized Stokes’ theorem that links a manifold to its boundary through differential forms. In short, the immediate takeaway is that topology provides the metric-free language needed to talk about calculus on surfaces and beyond.

Cornell Notes

The course starts with why calculus on surfaces requires new foundations: constraints can force motion to stay on a curved set like the sphere S², where standard calculus on open regions doesn’t directly apply. It then builds that foundation by introducing topology as a metric-free way to decide which subsets are “open.” In a metric space, openness comes from ε-balls, but topology abstracts this into axioms about a collection T of subsets of X. A topology must include ∅ and X, be closed under finite intersections, and be closed under arbitrary unions. Two edge-case examples—indiscrete (only ∅ and X) and discrete (all subsets)—show how the axioms work even without geometric distance.

Why does calculus on a sphere require more than the usual Euclidean notion of open sets?

On a surface like the 2-sphere S² (the boundary of a ball), motion or constraints restrict everything to the surface itself. Standard calculus tools are typically formulated for open domains in Euclidean space, where points can move in all nearby directions. Once the domain is a surface, “nearby points” must be understood intrinsically on the surface, not via ambient open sets in the surrounding space. That motivates replacing distance-based definitions with a more general framework for neighborhoods and openness.

How does topology remove the need for a metric?

In a metric space, openness is defined using ε-balls: a set is open if around each point there is a neighborhood of radius ε entirely contained in the set. Topology keeps the set X but replaces the distance function with a chosen collection T of subsets declared “open.” The only requirements are three axioms: ∅ and X are in T; the intersection of two open sets is open; and the union of any indexed family of open sets is open. This means openness is determined by T, not by a specific distance d.

What are the three axioms that define a topology on a set X?

A topology T on X satisfies: (1) ∅ ∈ T and X ∈ T; (2) if A ∈ T and B ∈ T, then A ∩ B ∈ T; (3) for any index set I and any family {A_i}_{i∈I} with A_i ∈ T, the union ⋃_{i∈I} A_i is in T. These axioms encode how “open sets” behave without referencing distances.

What distinguishes the indiscrete and discrete topologies?

The indiscrete topology is the minimal choice: T = {∅, X}. It satisfies the axioms because intersections and unions of these two sets stay within the same collection, but it makes almost everything non-open. The discrete topology is the maximal choice: T = P(X), the power set of X. Here every subset is open, so the axioms hold trivially. Both are edge cases that clarify the role of the topology’s axioms.

How do open sets become relative to the chosen topology?

Once openness is defined as membership in T, the statement “a set is open” depends on which topology T is chosen. Two different topologies on the same underlying set X can disagree about whether a particular subset counts as open, even though both satisfy the topology axioms.

Review Questions

What problem arises when constraints force motion to remain on a surface like S², and how does that motivate topology?
List the three axioms of a topology and explain how they correspond to closure properties of open sets.
Compare the indiscrete and discrete topologies on the same set X: what sets are open in each, and why do both satisfy the axioms?

Key Points

1
The course motivates topology by the need to extend calculus from open regions to curved surfaces such as the sphere S², where constraints restrict motion to a surface.
2
Topology abstracts away metric distance by defining openness through a chosen collection T of subsets of X rather than ε-balls.
3
A topology T on X must include ∅ and X, be closed under finite intersections, and be closed under arbitrary unions.
4
In topology, “open” is not an intrinsic geometric property of subsets; it is defined relative to the selected topology T.
5
Indiscrete topology (T = {∅, X}) and discrete topology (T = P(X)) are extreme examples that both satisfy the axioms but offer very different flexibility.
6
The long-term plan is to use topology as groundwork for differentiable manifolds, differential forms, and a generalized Stokes’ theorem connecting a manifold to its boundary.

Highlights

The sphere S² is used as the flagship example of a surface where calculus must be reformulated because standard open-domain assumptions fail.

Topology replaces distance-based neighborhoods with axioms about which subsets are declared open.

A topology is defined purely by three closure rules: include ∅ and X, close under intersections, and close under arbitrary unions.

Indiscrete and discrete topologies demonstrate how the same axioms can produce radically different notions of openness.

Topics

Manifolds
Topology
Open Sets
Differentiable Manifolds
Generalized Stokes' Theorem