The Central Idea

Geometry asks:

Topology asks:

Geometry is usually more detailed. Topology keeps the structural information that survives continuous deformation.

Topology Forgets Some Geometric Information

A homeomorphism may stretch, bend, compress, or distort a space.

Therefore quantities such as

are generally not preserved by homeomorphism.

Geometry Adds Structure to Topology

A topological space tells us which sets are open.

A geometric space often carries additional information, such as a metric, inner product, or curvature rule.

Thus one may think schematically:

\[ \boxed{ \text{geometric structure} = \text{topological structure} + \text{additional measurement data} }. \]

Metric Spaces Connect Geometry and Topology

A metric space is a set \(X\) equipped with a distance function

\[ d:X\times X\to[0,\infty). \]

The metric must satisfy:

\[ d(x,y)\ge0, \] \[ d(x,y)=0\iff x=y, \] \[ d(x,y)=d(y,x), \]

and the triangle inequality

\[ d(x,z)\le d(x,y)+d(y,z). \]

A Metric Generates Open Balls

Once a metric is given, we can define open balls:

\[ B_r(x)=\{y\in X:d(x,y) These open balls generate a topology.

A set \(U\subseteq X\) is open if every point \(x\in U\) lies in some ball contained in \(U\):

\[ \boxed{ x\in U \Longrightarrow \exists r>0 \text{ such that } B_r(x)\subseteq U }. \]

From Distance to Topology

The logical direction is important:

\[ \boxed{ \text{metric} \longrightarrow \text{open balls} \longrightarrow \text{topology} }. \]

Distance gives a precise geometric notion of closeness, and that notion of closeness determines which sets are open.

Topology Does Not Always Come from a Metric

Every metric generates a topology.

But not every topological space is metrizable.

A topological space is metrizable if some metric generates exactly its topology.

Thus topology is more general than metric geometry.

Different Metrics Can Generate the Same Topology

A topology does not necessarily determine a unique metric.

Two different distance functions can produce the same open sets.

When that happens, the metrics differ geometrically but describe the same underlying topology.

Example: Rescaling a Metric

Suppose \(d\) is a metric and define

\[ d'(x,y)=2d(x,y). \]

Distances have doubled, but an open ball in \(d'\) satisfies

\[ B^{d'}_r(x) = B^d_{r/2}(x). \]

Therefore \(d\) and \(d'\) generate the same topology.

Equivalent Metrics

Metrics that generate the same topology are called topologically equivalent metrics.

They may assign different numerical distances while agreeing on:

Homeomorphism vs. Isometry

A homeomorphism preserves topology.

An isometry preserves distance exactly.

If

\[ f:X\to Y \]

is an isometry, then

\[ d_Y(f(x),f(y))=d_X(x,y) \]

for all \(x,y\in X\).

Isometry Is Stronger than Homeomorphism

In ordinary metric-space settings, a bijective isometry is a homeomorphism.

Therefore,

\[ \boxed{ \text{isometric} \Longrightarrow \text{homeomorphic} }. \]

But the converse is false.

\[ \boxed{ \text{homeomorphic} \centernot\Longrightarrow \text{isometric} }. \]

Example: Intervals of Different Length

The intervals

\[ [0,1] \]

and

\[ [0,10] \]

are homeomorphic.

But with their usual Euclidean metrics they are not isometric because their diameters differ:

\[ \operatorname{diam}[0,1]=1, \qquad \operatorname{diam}[0,10]=10. \]

Geometric Invariants vs. Topological Invariants

A geometric invariant is preserved under a chosen class of geometric transformations.

A topological invariant is preserved under all homeomorphisms.

For example:

A Hierarchy of Structure

Different equivalence notions preserve different amounts of information.

A useful hierarchy is:

\[ \boxed{ \text{isometry} \Longrightarrow \text{diffeomorphism} \Longrightarrow \text{homeomorphism} } \]

in standard smooth geometric settings.

Each step forgets some structure.

Homeomorphism Forgets Measurement

A homeomorphism preserves continuity and open-set structure but may distort measurement.

Therefore a circle and an ellipse are homeomorphic even though their curvature and arc-length behavior differ.

Diffeomorphism Preserves Smooth Structure

A diffeomorphism is a smooth bijection with a smooth inverse.

It preserves the smooth-manifold structure.

Diffeomorphisms are more restrictive than homeomorphisms because continuity alone does not require differentiability.

Geometry on Manifolds

A topological manifold is locally homeomorphic to Euclidean space.

A smooth manifold adds differentiable structure.

A Riemannian manifold adds a smoothly varying inner product on tangent spaces.

This produces the hierarchy

\[ \boxed{ \text{topological manifold} \to \text{smooth manifold} \to \text{Riemannian manifold} }. \]

Riemannian Geometry Adds Measurement

A Riemannian metric allows us to measure:

All of this is additional structure beyond topology.

Topology Can Stay Fixed While Geometry Changes

A sphere can be equipped with many different Riemannian metrics.

The resulting surfaces may have different curvature distributions and distance functions.

Yet the underlying topological space remains

\[ S^2. \]

Example: Sphere and Ellipsoid

A sphere and an ellipsoid are homeomorphic.

Their local metric geometry differs:

But they share the same topology:

\[ g=0, \qquad \chi=2. \]

Curvature Is Geometric

Curvature measures how geometric structure departs from flatness.

It depends on the chosen metric.

A homeomorphism need not preserve curvature.

Therefore curvature is generally not a topological invariant.

Intrinsic Curvature

Some curvature can be measured entirely from distances and angles within a surface, without referring to how the surface sits in surrounding space.

Such curvature is intrinsic.

Gaussian curvature is the central example for two-dimensional Riemannian surfaces.

Extrinsic Curvature

A surface embedded in a larger space can also bend within that ambient space.

Such bending is extrinsic.

A sheet of paper can be rolled into a cylinder without stretching.

Its extrinsic appearance changes, while its intrinsic flat geometry remains locally unchanged.

Plane and Cylinder: Same Intrinsic Geometry Locally

A cylinder can be formed from a rectangular sheet without stretching the sheet.

Locally, distances and angles on the sheet are preserved.

The cylinder bends in three-dimensional space, but its Gaussian curvature remains

\[ \boxed{0}. \]

Gauss's Theorema Egregium

Gauss's Theorema Egregium states that Gaussian curvature is determined intrinsically by the metric of the surface.

This was a profound discovery:

certain curvature information can be detected entirely from measurements made within the surface.

Topology Constrains Curvature Globally

Although curvature itself is geometric, topology can impose global restrictions on curvature.

The most famous example for closed surfaces is the Gauss-Bonnet theorem.

Gauss-Bonnet Preview

For a closed oriented Riemannian surface \(M\),

\[ \boxed{ \int_M K\,dA = 2\pi\chi(M) }. \]

Here:

This equation directly connects geometry and topology.

Why Gauss-Bonnet Is Remarkable

The left side is geometric:

\[ \int_M K\,dA. \]

The right side is topological:

\[ 2\pi\chi(M). \]

Therefore a global integral of curvature is forced by the topology of the surface.

Sphere and Total Curvature

For the sphere,

\[ \chi(S^2)=2. \]

Gauss-Bonnet gives

\[ \int_{S^2}K\,dA = 2\pi(2) \] \[ \boxed{4\pi}. \]

The curvature may be redistributed under different metrics, but the total remains constrained by topology.

Torus and Total Curvature

For the torus,

\[ \chi(T^2)=0. \]

Therefore,

\[ \int_{T^2}K\,dA = 0. \]

A standard embedded torus has regions of positive and negative Gaussian curvature whose total balances to zero.

Genus and Curvature

For a connected closed orientable surface,

\[ \chi=2-2g. \]

Therefore Gauss-Bonnet becomes

\[ \int_M K\,dA = 2\pi(2-2g). \]

The topology of handles constrains the total curvature.

Geodesics

A geodesic is a curve that is locally straight according to the geometry of the space.

In Euclidean space, geodesics are straight lines.

On a sphere, geodesics are great circles.

Geodesics depend on geometry, not topology alone.

Homeomorphisms Need Not Preserve Geodesics

A homeomorphism may distort a shortest or locally straight path into a curve that is no longer geodesic.

Therefore geodesic structure is not purely topological.

Distance on Curved Spaces

On a Riemannian manifold, distance can be defined using curve lengths.

For points \(p,q\),

\[ d(p,q) = \inf \{ L(\gamma): \gamma\text{ joins }p\text{ to }q \}. \]

This metric then generates the manifold's topology under standard conditions.

Geometry Often Recovers the Topology

Once a Riemannian metric is placed on a manifold, its distance function generates the original manifold topology.

Thus the geometric and topological structures are compatible.

Geometry refines rather than replaces the topology.

Euclidean Geometry and Euclidean Topology

The usual Euclidean metric on

\[ \mathbb{R}^n \]

is

\[ d(x,y) = \sqrt{ (x_1-y_1)^2+\cdots+(x_n-y_n)^2 }. \]

Its open balls generate the standard Euclidean topology.

Other Norms Give the Same Topology on \(\mathbb{R}^n\)

On finite-dimensional Euclidean space, many different norms generate the same topology.

Examples include:

\[ \|x\|_2 = \sqrt{x_1^2+\cdots+x_n^2}, \] \[ \|x\|_1 = |x_1|+\cdots+|x_n|, \]

and

\[ \|x\|_\infty = \max_i|x_i|. \]

Different Unit Balls, Same Topology

Under different norms, metric balls have different geometric shapes.

In \(\mathbb{R}^2\):

Yet all of these norms generate the same standard topology on \(\mathbb{R}^2\).

Taxicab Geometry and Topology

The taxicab metric on \(\mathbb{R}^2\) is

\[ d_1((x_1,y_1),(x_2,y_2)) = |x_1-x_2|+|y_1-y_2|. \]

Its geometry differs from Euclidean geometry.

But it generates the same topology.

Thus geometric notions such as shortest paths can change while continuity and openness remain the same.

Non-Euclidean Geometry Can Share Familiar Topology

A sphere carries non-Euclidean spherical geometry.

Yet topologically it remains the two-sphere

\[ S^2. \]

Hyperbolic surfaces can also carry rich non-Euclidean geometry while having familiar topological classifications by genus.

Same Topology, Different Geometries

A single topological surface can support different geometric structures.

For example, a torus can be represented with:

The topology remains that of a torus.

Topology Can Limit Possible Geometries

Not every geometric structure can exist on every topological space.

Global topology may force restrictions involving:

The Hairy Ball Theorem

A famous topological result says that there is no continuous nowhere-zero tangent vector field on an even-dimensional sphere such as

\[ S^2. \]

Informally, one cannot comb a hairy sphere flat without creating at least one singular point.

This is another example of topology constraining geometric structure.

Embeddings Connect Intrinsic and Extrinsic Geometry

An embedding places a topological space inside another space while preserving its intrinsic topology.

Once embedded, new extrinsic questions become available:

Intrinsic vs. Extrinsic Questions

An intrinsic question can be answered using only measurements made within the space.

An extrinsic question depends on the ambient space.

For example:

Knot Theory Shows Why Embedding Matters

Every knot is intrinsically homeomorphic to

\[ S^1. \]

The unknot and trefoil differ because of how that circle is embedded in three-dimensional space.

Thus topology itself includes both intrinsic spaces and the topology of embeddings.

Surface Classification Is Topological, Not Geometric

Connected compact surfaces are classified by data such as:

Exact geometric shape does not enter the classification.

Geometry Can Refine a Topological Classification

Once the topological type is known, geometry asks finer questions.

For a torus, for example, one can then ask:

Topology Is Often the Coarse First Layer

A useful problem-solving order is:

\[ \boxed{ \text{topological type} \longrightarrow \text{smooth structure} \longrightarrow \text{geometric structure} }. \]

First determine what the space is topologically.

Then study finer geometric data.

Geometry Can Reveal Topology

The relationship also works in the opposite direction.

Geometric measurements can reveal topological information.

Gauss-Bonnet is a central example:

\[ \int_MK\,dA=2\pi\chi(M). \]

Measuring total curvature can determine Euler characteristic.

Topology Can Reveal Impossible Geometries

If a proposed geometric structure would force topological invariants to take impossible values, then that geometry cannot exist globally.

This principle appears throughout differential geometry and global analysis.

Compactness Has Geometric Consequences

Compactness is topological, but it often combines with geometric structure to produce strong results.

On compact metric or Riemannian spaces, one often obtains:

Connectedness Has Geometric Consequences

Connectedness is topological.

In geometric settings, it often supports statements about:

Geometry can sharpen these consequences.

Convexity Is Geometric, Not Topological

A subset \(C\subseteq\mathbb{R}^n\) is convex if every line segment between two points of \(C\) lies inside \(C\).

Convexity depends on the linear geometry of the ambient space.

A homeomorphism can carry a convex disk to a highly distorted nonconvex region.

Therefore convexity is not a topological invariant.

But Convexity Implies Topological Properties

Although convexity is not topological, it has topological consequences.

Every convex subset of \(\mathbb{R}^n\) is path connected.

In fact, every convex set is contractible.

Thus geometric hypotheses can force topological structure.

Contractibility

A space is contractible if it can be continuously shrunk to a point within itself.

Convex subsets of Euclidean space are contractible using the straight-line homotopy

\[ H(x,t)=(1-t)x+tp, \]

where \(p\) is a fixed point in the convex set.

Geometry and Topology Use Different Equivalence Relations

Several notions of "same" may appear:

Each preserves a different level of structure.

Congruence

Congruence preserves Euclidean size and shape.

Congruent objects are related by rigid motions such as translations, rotations, and reflections.

This is much stronger than homeomorphism.

Similarity

Similarity allows uniform scaling while preserving angles and shape ratios.

Similar figures are homeomorphic, but many homeomorphic spaces are not similar.

Homeomorphism

Homeomorphism preserves topological structure but allows broad geometric distortion.

A circle and square boundary are homeomorphic.

A sphere and ellipsoid are homeomorphic.

A sphere and torus are not.

Homotopy Equivalence Is Coarser Still

Homotopy equivalence ignores more structure than homeomorphism.

A disk is not homeomorphic to a point, but it is homotopy equivalent to a point because it is contractible.

This introduces another layer between exact topology and broad deformation behavior.

Common Error: Thinking Topology Ignores All Shape

Topology ignores exact metric shape, but it still detects structural shape.

It distinguishes:

Common Error: Thinking Geometry and Topology Are Unrelated

Geometry frequently generates topology through metrics.

Topology constrains global geometric behavior.

The Gauss-Bonnet theorem is one of the clearest examples of the two subjects interacting directly.

Common Error: Thinking Curvature Is Topological

Curvature depends on geometric structure.

Homeomorphic spaces may carry very different curvature.

What topology can control are certain global combinations of curvature under additional geometric hypotheses.

Common Error: Thinking Every Topology Comes from a Metric

Every metric space is a topological space.

But some topological spaces are not metrizable.

Therefore metric geometry is an important special source of topologies, not the entire subject.

Common Error: Thinking Same Topology Means Same Distances

Different metrics may generate the same topology.

Topological equivalence remembers open-set structure, not exact numerical distances.

Common Error: Confusing Intrinsic and Extrinsic Curvature

A cylinder bends in \(\mathbb{R}^3\), but intrinsically it is flat.

Its Gaussian curvature is

\[ 0. \]

Visible bending in ambient space is not the same as intrinsic curvature.

Common Error: Thinking Homeomorphism Preserves Shortest Paths

Shortest paths depend on a metric.

A homeomorphism may dramatically distort lengths.

Therefore geodesics are generally not preserved by arbitrary homeomorphism.

A Reliable Procedure for Separating Geometric and Topological Questions

  1. Identify the space and its topology.
  2. Identify any extra structure: metric, smooth structure, embedding, or curvature.
  3. Ask whether the property depends on measurement.
  4. If it survives all homeomorphisms, treat it as topological.
  5. If it depends on distance, angle, or curvature, treat it as geometric.
  6. Look for theorems that connect the two levels, such as Gauss-Bonnet.

A Reliable Procedure for Metric-Induced Topology

  1. Start with the metric \(d\).
  2. Construct open balls \(B_r(x)\).
  3. Declare a set open when every point contains some ball lying inside it.
  4. Use those open sets to define continuity, neighborhoods, closure, and compactness.
  5. Compare different metrics by checking whether they generate the same open sets.

Example: Circle vs. Ellipse

A circle and ellipse are homeomorphic.

They are both compact, connected one-dimensional manifolds without boundary.

Their curvature differs geometrically.

Therefore:

\[ \boxed{ \text{same topology} \quad\text{but}\quad \text{different geometry} }. \]

Example: Sphere vs. Torus

Both are smooth compact surfaces.

But

\[ \chi(S^2)=2 \]

and

\[ \chi(T^2)=0. \]

Therefore they have different topology regardless of how their geometry is changed.

Example: Euclidean vs. Taxicab Plane

The Euclidean and taxicab metrics assign different distances.

Their metric balls have different shapes.

Yet they generate the same topology on

\[ \mathbb{R}^2. \]

Therefore continuity and connectedness agree even though shortest-path geometry differs.

Practice 1: Metric to Topology

Complete:

\[ \text{metric} \longrightarrow \boxed{\phantom{\text{open balls}}} \longrightarrow \text{topology}. \]

Practice 2: Homeomorphism vs. Isometry

Which is stronger: homeomorphism or isometry?

Practice 3: Boundedness

Is boundedness a topological invariant?

Practice 4: Curvature

Is Gaussian curvature generally preserved by arbitrary homeomorphism?

Practice 5: Same Topology, Different Geometry

Give one example of two homeomorphic spaces that can have different geometry.

Practice 6: Gauss-Bonnet

Complete the closed-surface formula:

\[ \int_M K\,dA = \boxed{\phantom{2\pi\chi(M)}}. \]

Practice 7: Sphere

If

\[ \chi(S^2)=2, \]

what total curvature does Gauss-Bonnet give?

Practice 8: Torus

If

\[ \chi(T^2)=0, \]

what is

\[ \int_{T^2}K\,dA? \]

Practice 9: Convexity

Is convexity a topological invariant?

Practice 10: Metrizability

Does every topological space arise from a metric?

Practice Answers

Practice 1

\[ \boxed{\text{open balls}}. \]

Practice 2

Isometry is stronger because it preserves exact distances.

\[ \boxed{ \text{isometry} \Longrightarrow \text{homeomorphism} } \]

for bijective isometries in ordinary metric settings.

Practice 3

No. For example,

\[ (0,1)\cong\mathbb{R}, \]

but only the first is bounded.

Practice 4

No. Curvature depends on geometric structure and can change under a homeomorphism.

Practice 5

A sphere and an ellipsoid are homeomorphic but can have different metric and curvature data.

A circle and ellipse provide another example.

Practice 6

\[ \boxed{2\pi\chi(M)}. \]

Practice 7

\[ \int_{S^2}K\,dA = 2\pi(2) \] \[ \boxed{4\pi}. \]

Practice 8

\[ \boxed{0}. \]

Practice 9

No. A homeomorphism can distort a convex set into a nonconvex one.

Practice 10

No. Some topological spaces are not metrizable.

What to Remember

The central relationship is:

\[ \boxed{ \text{geometry refines topology, while topology constrains geometry} }. \]

Where We Go Next

Geometry and topology are not competing descriptions of space. They operate at different levels. Topology identifies the broad structural type, while geometry supplies measurement, curvature, geodesics, and quantitative form.

From here, the natural next steps include manifolds, differential geometry, Riemannian metrics, curvature, Gauss-Bonnet, homotopy, algebraic topology, and geometric topology. These subjects study spaces by deliberately combining topological structure with increasingly rich geometric information.