The Central Idea

A manifold can be summarized by one guiding idea:

Locally familiar, globally flexible.

Near any one point, a manifold behaves like ordinary Euclidean space of a fixed dimension.

Globally, however, the manifold may curve, close onto itself, contain holes, or have a shape that cannot be represented by one flat coordinate picture.

What Does Locally Mean?

In geometry, local means "within a sufficiently small neighborhood of a point."

The entire object does not need to resemble Euclidean space.

Only a small region around each point must do so.

This local viewpoint is what allows a curved surface such as a sphere to qualify as a two-dimensional manifold.

What Does Euclidean Mean Here?

A space is locally Euclidean of dimension \(n\) when every point has a neighborhood that can be described by coordinates in an open region of

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

The number \(n\) is the dimension of the manifold.

Dimension

The dimension of a manifold is the number of independent local coordinates needed to describe points near any location.

Examples:

Curves as One-Dimensional Manifolds

A smooth curve can often be described locally by one parameter.

For example, a circle can be written

\[ x=\cos t,\qquad y=\sin t. \]

Near any chosen point, changing \(t\) slightly moves along the circle in one independent direction.

Therefore the circle is a one-dimensional manifold.

A Circle Is Not a Line Globally

A circle is locally similar to a line, but globally it is different.

A line extends indefinitely in two directions.

A circle closes onto itself.

This is our first important example of local similarity without global equivalence.

Surfaces as Two-Dimensional Manifolds

A smooth surface is locally described by two coordinates.

Near a point on a sphere, for example, we can use two coordinates to specify nearby positions.

These might be

Because two local coordinates are needed, the sphere is two-dimensional.

Embedded Dimension Is Different

A sphere is commonly drawn inside three-dimensional space:

\[ x^2+y^2+z^2=R^2. \]

Yet the sphere itself is two-dimensional.

The distinction is:

A two-dimensional manifold may be embedded in \(\mathbb{R}^3\), but that does not make it three-dimensional.

Neighborhoods

A neighborhood of a point is a region containing that point together with nearby points.

For a manifold, every point must have a neighborhood that can be represented by ordinary Euclidean coordinates.

This neighborhood may be very small.

Coordinate Charts

A coordinate chart assigns Euclidean coordinates to points in a local region of a manifold.

A chart consists conceptually of:

We write

\[ \varphi:U\to\mathbb{R}^n. \]

Why Charts Matter

Charts let us replace a small curved region by coordinates in ordinary Euclidean space.

Once coordinates are available, we can use familiar tools such as

Maps and Globes

Geographic maps provide an excellent intuition for charts.

The Earth is approximately spherical, but a regional map converts a portion of the surface into planar coordinates.

No single ordinary flat map can represent the entire sphere without distortion or singular behavior.

Multiple maps are therefore needed.

A manifold is handled in much the same way.

An Atlas

A collection of coordinate charts covering a manifold is called an atlas.

The term is intentionally similar to a geographic atlas.

Different charts describe different regions, and together they cover the entire space.

Overlapping Charts

Two coordinate charts may overlap.

In the overlapping region, a point receives two coordinate descriptions.

We therefore need a way to translate from one coordinate system to the other.

These translations are called transition maps.

Transition Maps

Suppose two charts are

\[ \varphi:U\to\mathbb{R}^n \]

and

\[ \psi:V\to\mathbb{R}^n. \]

On the overlap \(U\cap V\), the coordinate change is

\[ \psi\circ\varphi^{-1}. \]

This map converts \(\varphi\)-coordinates into \(\psi\)-coordinates.

Smooth Manifolds

A smooth manifold is a manifold whose overlapping coordinate charts fit together smoothly.

Informally, this means the transition maps can be differentiated repeatedly without producing abrupt breaks or corners.

Smoothness allows calculus to be carried from Euclidean space onto the manifold.

Topological Manifolds

Before adding smoothness, one can study topological manifolds.

These require local Euclidean structure and appropriate topological conditions but do not initially require differentiability.

Smooth manifolds add more structure on top of the underlying topological manifold.

Why the Distinction Matters

Different questions require different levels of structure.

A manifold is therefore not one single rigid concept but a foundation on which additional geometric structure can be built.

The Real Line

The simplest one-dimensional manifold is

\[ \mathbb{R}. \]

Every point already has an ordinary real coordinate.

No special atlas construction is required because the entire space can be covered by one global chart.

Euclidean Space

More generally,

\[ \mathbb{R}^n \]

is the basic example of an \(n\)-dimensional manifold.

Its standard coordinates are

\[ (x_1,x_2,\ldots,x_n). \]

Every manifold is locally modeled on some \(\mathbb{R}^n\).

The Circle \(S^1\)

The unit circle is

\[ S^1=\{(x,y)\in\mathbb{R}^2:x^2+y^2=1\}. \]

It is a one-dimensional manifold embedded in \(\mathbb{R}^2\).

One parameter is enough locally, but no single ordinary angular coordinate gives a globally seamless real-number description without choosing a cut.

The Sphere \(S^2\)

The unit sphere is

\[ S^2=\{(x,y,z)\in\mathbb{R}^3:x^2+y^2+z^2=1\}. \]

It is a two-dimensional manifold embedded in \(\mathbb{R}^3\).

Local coordinates require two numbers.

Why Latitude and Longitude Are Not One Perfect Global Chart

Latitude and longitude work over much of the sphere.

But longitude becomes undefined at the poles.

There is also a coordinate seam where longitudes wrap from one conventional endpoint to another.

The sphere itself is smooth there; the problem belongs to the coordinate system, not the manifold.

Coordinate Singularities

A coordinate singularity occurs when a particular coordinate system breaks down even though the underlying space remains regular.

Longitude at the poles is a familiar example.

This distinction becomes important throughout geometry and physics.

Stereographic Projection

Another useful coordinate system for the sphere is stereographic projection.

Choose one point of the sphere as a projection pole and project the remaining sphere onto a plane.

This gives a chart covering every point except the chosen pole.

A second stereographic chart from the opposite pole can cover the missing point.

The Torus

A torus is the familiar doughnut-shaped surface.

It is a two-dimensional manifold.

Locally, a small patch looks like a region of the plane.

Globally, however, a torus differs dramatically from a sphere because it has a hole.

Local Geometry Does Not Determine Global Topology

A sphere and a torus are both locally two-dimensional.

That fact alone does not make them globally equivalent.

Their global topology differs.

This is another central theme:

Local dimension does not determine global shape.

The Cylinder

A circular cylinder is also a two-dimensional manifold.

A small patch can be flattened without stretching, reflecting the fact that the cylinder has zero Gaussian curvature.

Globally, however, one direction wraps around periodically.

Open Sets

In the formal definition of a manifold, local coordinate neighborhoods are matched with open sets of Euclidean space.

Intuitively, an open set contains a little room around each of its points.

This prevents ordinary interior points from behaving like artificial edges created only by the coordinate representation.

Manifolds Without Boundary

In the simplest manifold definition used in this lesson, every point has a neighborhood modeled on an open subset of

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

Such manifolds have no boundary.

The circle, sphere, torus, and full Euclidean spaces are examples.

Manifolds with Boundary

Geometry also allows manifolds with boundary.

Near a boundary point, the local model resembles a half-space rather than a full Euclidean neighborhood.

In \(n\) dimensions, the standard half-space is

\[ \mathbb{H}^n=\{(x_1,\ldots,x_n)\in\mathbb{R}^n:x_n\ge0\}. \]

An Interval as a Manifold with Boundary

Consider the closed interval

\[ [0,1]. \]

Interior points look locally like ordinary pieces of the real line.

The endpoints \(0\) and \(1\) look locally like half-lines.

Thus \([0,1]\) is a one-dimensional manifold with boundary.

A Disk as a Manifold with Boundary

The closed disk

\[ D^2=\{(x,y):x^2+y^2\le1\} \]

is a two-dimensional manifold with boundary.

Its boundary is the circle

\[ S^1. \]

Dimension of the Boundary

For an \(n\)-dimensional manifold with boundary, the boundary—when nonempty—is typically an \((n-1)\)-dimensional manifold.

Examples:

Not Every Shape Is a Manifold

A space fails to be a manifold when some point does not have the required local Euclidean form.

A common example is a crossing of two line segments.

Away from the crossing, the space looks locally like a line.

At the crossing, however, four branches meet.

That neighborhood does not look like an ordinary open interval.

Branch Points

A branch point can prevent a one-dimensional space from being a manifold.

Consider the letter \(Y\) as a geometric graph.

Most points lie on ordinary line-like branches.

At the central junction, three branches meet.

The junction does not have a neighborhood homeomorphic to an open interval or half-interval.

Cones and Singular Points

A cone can illustrate the idea of a singular point.

Away from the apex, a cone behaves like a smooth two-dimensional surface.

At the apex, smooth tangent behavior fails in the ordinary embedded sense.

Depending on which manifold structure is being studied, that apex may be treated as a singularity rather than an ordinary smooth point.

Local Coordinates on a Surface

A surface patch can be parameterized by two variables:

\[ \mathbf{r}(u,v) = \bigl(x(u,v),y(u,v),z(u,v)\bigr). \]

The variables \(u\) and \(v\) act as local coordinates.

Nearby coordinate pairs correspond to nearby points on the surface.

Tangent Vectors from Coordinates

Differentiate a parameterization with respect to its coordinates:

\[ \mathbf{r}_u=\frac{\partial\mathbf{r}}{\partial u}, \] \[ \mathbf{r}_v=\frac{\partial\mathbf{r}}{\partial v}. \]

These vectors point along coordinate directions on the surface.

When they are linearly independent, they span the local tangent plane.

The Tangent Space

At a point \(p\) on an \(n\)-dimensional smooth manifold, the tangent space is an \(n\)-dimensional vector space denoted

\[ T_pM. \]

It represents the possible instantaneous directions of motion through \(p\) while remaining on the manifold.

Tangent Space of a Curve

For a smooth one-dimensional curve, the tangent space at a regular point is a one-dimensional vector space.

If

\[ \mathbf{r}(t) \]

parameterizes the curve, then

\[ \mathbf{r}'(t) \]

provides a tangent direction when it is nonzero.

Tangent Space of a Surface

For a smooth two-dimensional surface, the tangent space is a plane-like two-dimensional vector space.

With local coordinates \(u,v\), it is spanned by

\[ \mathbf{r}_u,\qquad\mathbf{r}_v. \]

The tangent plane is the best first-order linear approximation to the surface near the point.

Why Tangent Spaces Matter

The manifold itself may be curved, but its tangent space is linear.

This lets us bring linear algebra into curved geometry.

Tangent spaces support:

Velocity on a Manifold

Suppose a curve on a manifold is written

\[ \gamma(t). \]

Its velocity at time \(t\) is

\[ \gamma'(t). \]

At the point

\[ p=\gamma(t), \]

this velocity belongs to the tangent space

\[ T_pM. \]

Functions on Manifolds

A real-valued function on a manifold has the form

\[ f:M\to\mathbb{R}. \]

In a coordinate chart, the function can be rewritten as an ordinary multivariable function of local coordinates.

This is how familiar calculus enters the manifold setting.

Example: Height on a Sphere

On the unit sphere, define

\[ f(x,y,z)=z. \]

This function measures vertical height.

In local coordinates on the sphere, \(f\) becomes a function of two local variables rather than three independent ambient variables.

Intrinsic and Extrinsic Viewpoints

A manifold can be studied

This distinction appeared earlier in the lessons on curvature and geodesic distance.

The Sphere Intrinsically

A two-dimensional traveler restricted to a sphere can study

without referring to the center of the sphere or the surrounding three-dimensional space.

Riemannian Metrics

A smooth manifold by itself does not automatically specify lengths or angles.

To measure them, we can add a Riemannian metric.

At each point, the metric provides an inner product on the tangent space.

This lets us measure lengths and angles of tangent vectors.

Local Metric Form

In local coordinates

\[ (x^1,\ldots,x^n), \]

a Riemannian metric can be written

\[ ds^2 = \sum_{i,j=1}^{n} g_{ij}\,dx^i\,dx^j. \]

The coefficients \(g_{ij}\) may vary from point to point.

This formula generalizes the familiar Euclidean distance element.

Euclidean Metric as a Special Case

In ordinary Cartesian coordinates,

\[ ds^2=dx^2+dy^2 \]

in the plane, and

\[ ds^2=dx^2+dy^2+dz^2 \]

in three-dimensional Euclidean space.

Curved manifolds may have coordinate-dependent metric coefficients.

The Sphere's Metric

On a sphere of radius \(R\), using colatitude \(\theta\) and longitude \(\phi\), the surface metric is

\[ ds^2 = R^2\,d\theta^2 + R^2\sin^2\theta\,d\phi^2. \]

This formula describes distances entirely within the spherical surface.

Geodesics Return

Once a Riemannian metric is defined, path length can be measured.

Geodesics can then be characterized as locally extremal or locally straight paths with respect to that metric.

On a sphere, the geodesics are great circles.

On a general manifold, geodesics are determined by the local metric.

Curvature Returns

Curvature can also be defined intrinsically on a Riemannian manifold.

Earlier lessons compared positive, zero, and negative curvature.

Manifold language now provides the broader setting in which those ideas can be developed systematically.

Local Flatness Does Not Mean Zero Curvature

Saying that a manifold is locally Euclidean does not mean that it has zero curvature.

Local Euclidean structure means that coordinates exist.

It does not mean the metric coefficients must equal the flat Euclidean metric throughout a neighborhood.

A sphere is locally Euclidean as a manifold but has positive Gaussian curvature.

Local Approximation

At sufficiently small scales, a smooth manifold can often be approximated by its tangent space.

This is analogous to approximating a smooth curve by its tangent line.

The approximation becomes more accurate as the neighborhood shrinks.

Global Properties

Manifolds may differ globally even when they have the same dimension and similar local geometry.

Global questions include:

Connectedness

A manifold is connected if it is all in one topological piece.

The circle and sphere are connected.

Two disjoint circles considered together form a disconnected one-dimensional manifold.

Compactness

Informally, a compact manifold is globally bounded in a topological sense and has no points "escaping to infinity."

The circle, sphere, and torus are compact.

The real line and Euclidean plane are not compact.

A precise definition belongs to topology, but the distinction is already geometrically useful.

Orientation

Some manifolds admit a consistent notion of orientation.

On an oriented surface, one can consistently distinguish two local orientation choices.

The sphere and torus are orientable.

The Möbius strip is the standard introductory example of a nonorientable surface with boundary.

The Möbius Strip

A Möbius strip can be constructed from a rectangular strip by giving one end a half twist before joining the ends.

Locally it looks like an ordinary two-dimensional surface.

Globally, following a local orientation around the strip reverses it.

This demonstrates once again that global behavior can differ from local behavior.

Higher-Dimensional Manifolds

Manifolds are not restricted to curves and surfaces.

An \(n\)-dimensional manifold is locally modeled on

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

Higher-dimensional manifolds cannot always be visualized directly, but their coordinate descriptions and mathematical properties remain precise.

The 3-Sphere

The three-dimensional sphere is

\[ S^3 = \left\{ (x_1,x_2,x_3,x_4)\in\mathbb{R}^4: x_1^2+x_2^2+x_3^2+x_4^2=1 \right\}. \]

It is a three-dimensional manifold embedded in four-dimensional Euclidean space.

Three local coordinates are required.

Configuration Spaces

A manifold can represent all possible states of a physical or mechanical system.

Such a space is often called a configuration space.

For example, the orientation of a rotating object may be represented by a manifold rather than by an ordinary Euclidean box of coordinates.

Robotics

In robotics, each joint may contribute one or more degrees of freedom.

The complete set of robot configurations can form a manifold.

Motion planning then becomes a problem of finding suitable paths through that configuration manifold.

Physics

Manifolds are central in modern physics.

In general relativity, spacetime is modeled as a four-dimensional manifold equipped with a metric structure appropriate to relativity.

Curvature then describes gravitational geometry.

The mathematics is far more advanced than this introduction, but the local coordinate idea is the same.

Data and Manifold Models

In data analysis and machine learning, high-dimensional observations may sometimes concentrate near a lower-dimensional structure.

This motivates manifold learning, which seeks lower-dimensional representations that preserve meaningful local relationships.

The mathematical setting is often approximate rather than an exact smooth manifold, but the geometric intuition is closely related.

Why One Coordinate System May Fail

A manifold may have perfectly regular geometry even though every attempted single coordinate system has a seam, pole, degeneracy, or missing point.

The solution is not to conclude that the manifold is defective.

The solution is to use multiple overlapping charts.

Common Error: Confusing a Manifold with Its Picture

A sphere drawn in three-dimensional space is not three-dimensional as a manifold.

The drawing shows an embedding.

The manifold dimension is determined by local coordinates on the object itself.

Common Error: Assuming Every Curved Object Is a Manifold

Curvature alone does not determine whether something is a manifold.

The essential question is local structure.

A branching junction or singular crossing can prevent a space from being locally Euclidean.

Common Error: Assuming Locally Euclidean Means Globally Euclidean

A circle is locally line-like but is not globally a line.

A sphere is locally plane-like but is not globally a plane.

Local Euclidean structure provides coordinates, not global flatness or global equivalence.

Common Error: Confusing Coordinate Failure with Geometric Failure

Latitude-longitude coordinates fail at the poles.

The sphere does not.

A coordinate singularity may reveal a limitation of the chart rather than a singularity of the manifold.

Common Error: Confusing Manifold Dimension with Ambient Dimension

The circle

\[ S^1\subset\mathbb{R}^2 \]

is one-dimensional.

The sphere

\[ S^2\subset\mathbb{R}^3 \]

is two-dimensional.

The ambient space may require more coordinates than the manifold itself.

Common Error: Assuming a Smooth Manifold Automatically Has a Distance

Smooth structure tells us how to perform calculus.

It does not by itself choose a specific measurement of lengths and angles.

A Riemannian metric supplies that additional information.

Common Error: Treating Local Coordinates as Unique

A manifold can have many valid coordinate systems.

The same point may be represented by different coordinate values in different charts.

Geometry should not depend on an arbitrary choice of coordinates.

A Reliable Procedure for Identifying a Manifold Dimension

  1. Choose a typical point.
  2. Examine a sufficiently small neighborhood.
  3. Determine how many independent coordinates are needed locally.
  4. Check that the same local dimension applies at every ordinary point.
  5. Distinguish intrinsic dimension from the dimension of any ambient space.

A Reliable Procedure for Thinking About Charts

  1. Choose a neighborhood \(U\) on the manifold.
  2. Assign local coordinates in \(\mathbb{R}^n\).
  3. Check that the coordinate representation is one-to-one on the chosen neighborhood.
  4. Use another chart where the first chart fails.
  5. Check that overlapping charts can be translated consistently.
  6. For smooth manifolds, require smooth transition maps.

A Reliable Procedure for Distinguishing Local and Global Questions

  1. Ask whether the property can be checked in a small neighborhood.
  2. If yes, treat it as primarily local.
  3. If the property depends on the entire space, treat it as global.
  4. Do not infer global topology from local dimension alone.
  5. Use charts for local calculations and broader topology for global structure.

Example: Is a Circle a Manifold?

Consider

\[ S^1=\{(x,y):x^2+y^2=1\}. \]

Near any point, a sufficiently short arc can be represented by one real coordinate.

Therefore each point has a neighborhood resembling an open interval.

Hence

\[ \boxed{S^1\text{ is a one-dimensional manifold}}. \]

Example: Is a Sphere a Manifold?

Consider

\[ S^2=\{(x,y,z):x^2+y^2+z^2=1\}. \]

Every point has a small surface neighborhood that can be represented using two coordinates.

Therefore

\[ \boxed{S^2\text{ is a two-dimensional manifold}}. \]

Example: Is a Crossed Pair of Lines a One-Dimensional Manifold?

Consider two lines crossing at the origin.

Away from the origin, each point has a line-like neighborhood.

At the origin, however, four local branches meet.

No sufficiently small neighborhood of the origin looks like an ordinary open interval.

Therefore the crossed space is

\[ \boxed{\text{not a one-dimensional manifold at the crossing}}. \]

Example: A Parameterized Surface

Consider the cylinder

\[ \mathbf{r}(u,v) = (\cos u,\sin u,v). \]

Differentiate:

\[ \mathbf{r}_u = (-\sin u,\cos u,0), \] \[ \mathbf{r}_v = (0,0,1). \]

These vectors are linearly independent everywhere.

Thus \(u\) and \(v\) provide local coordinates and span a two-dimensional tangent space.

Example: Tangent Plane to the Unit Sphere

Let

\[ p=(x_0,y_0,z_0) \]

lie on the unit sphere, so

\[ x_0^2+y_0^2+z_0^2=1. \]

The radius vector

\[ \mathbf{n}=p \]

is normal to the sphere at \(p\).

A tangent vector \(\mathbf{v}\) therefore satisfies

\[ p\cdot\mathbf{v}=0. \]

The tangent space is

\[ \boxed{ T_pS^2 = \{\mathbf{v}\in\mathbb{R}^3:p\cdot\mathbf{v}=0\} }. \]

Example: Dimension from Constraints

The sphere in \(\mathbb{R}^3\) is defined by one independent equation:

\[ x^2+y^2+z^2=1. \]

Away from singular situations, one independent smooth constraint reduces local dimension by one.

Thus the expected dimension is

\[ 3-1=2. \]

This agrees with the two-dimensional manifold structure of the sphere.

Example: A 3-Sphere Dimension Count

The 3-sphere lies in \(\mathbb{R}^4\) and satisfies

\[ x_1^2+x_2^2+x_3^2+x_4^2=1. \]

There are four ambient coordinates and one independent constraint.

Therefore its local dimension is

\[ \boxed{4-1=3}. \]

Practice 1: Dimension

What is the manifold dimension of each object?

  1. a circle;
  2. a sphere;
  3. a torus surface;
  4. ordinary Euclidean space \(\mathbb{R}^4\).

Practice 2: Intrinsic versus Ambient Dimension

A curve is embedded in \(\mathbb{R}^3\).

If one local parameter is sufficient to describe nearby points on the curve, what is its intrinsic dimension?

Practice 3: Local or Global?

Classify each feature as primarily local or global:

  1. the dimension near a point;
  2. whether a torus has a hole;
  3. the tangent plane at a point;
  4. whether a manifold is compact.

Practice 4: Coordinate Failure

Longitude is undefined at Earth's poles.

Does this mean the sphere fails to be a manifold at the poles? Explain.

Practice 5: Manifold or Not?

Consider the letter \(Y\) as three line segments meeting at one central point.

Is the entire figure a one-dimensional manifold?

Practice 6: Boundary

Is the closed disk

\[ x^2+y^2\le1 \]

a manifold without boundary or a manifold with boundary?

What is its boundary?

Practice 7: Tangent Dimension

If \(M\) is a smooth five-dimensional manifold, what is the dimension of

\[ T_pM \]

at an ordinary point \(p\)?

Practice 8: Sphere Constraint

The unit sphere is defined in \(\mathbb{R}^3\) by

\[ x^2+y^2+z^2=1. \]

Use the ambient-dimension-minus-independent-constraint idea to predict its dimension.

Practice 9: Smooth Structure and Metric Structure

Does specifying that a space is a smooth manifold automatically determine a unique formula for measuring lengths and angles?

Practice 10: Sphere and Plane

Explain how a sphere can be locally Euclidean and still have positive curvature.

Practice Answers

Practice 1

The dimensions are:

\[ \boxed{1,\ 2,\ 2,\ 4}. \]

A circle needs one local coordinate; a sphere and torus surface each need two; \(\mathbb{R}^4\) needs four.

Practice 2

The intrinsic dimension is

\[ \boxed{1}. \]

The fact that the curve sits inside \(\mathbb{R}^3\) does not change its one-dimensional local structure.

Practice 3

Practice 4

No.

The longitude coordinate breaks down at the poles, but another chart can be used there.

The failure belongs to the coordinate system rather than the sphere itself.

Practice 5

No.

At the central junction, three branches meet. A neighborhood of that point does not resemble an open interval or half-interval.

Therefore the full \(Y\)-shaped space is not a one-dimensional manifold at the branch point.

Practice 6

The disk is a

\[ \boxed{\text{two-dimensional manifold with boundary}}. \]

Its boundary is

\[ \boxed{x^2+y^2=1}, \]

the unit circle.

Practice 7

The tangent space has the same dimension as the manifold:

\[ \boxed{\dim T_pM=5}. \]

Practice 8

There are three ambient coordinates and one independent constraint:

\[ 3-1=2. \]

Therefore the sphere is

\[ \boxed{2\text{-dimensional}}. \]

Practice 9

No.

Smooth structure determines how differentiable coordinates fit together. A Riemannian metric is additional data that specifies lengths and angles.

Practice 10

Local Euclidean structure means that small neighborhoods admit coordinates in \(\mathbb{R}^2\).

It does not require the local metric to be globally identical to the flat Euclidean metric.

A sphere therefore can be a two-dimensional manifold while carrying positive intrinsic curvature.

What to Remember

Where We Go Next

The lessons in this section have explored several ways geometry can change: curvature can alter angle sums, metrics can redefine distance, projective geometry can replace parallelism, finite geometry can restrict the entire incidence system to finitely many objects, and manifolds can separate local Euclidean behavior from global shape.

The final lesson, Geometry as Assumptions and Consequences, brings these ideas together. We will examine how geometric systems arise from selected definitions and axioms, how changing assumptions changes the theorems that follow, and why Euclidean and non-Euclidean geometries are best understood as logically organized mathematical structures rather than as competing descriptions with only one possible starting point.