Unification — all three theorems as generalized Stokes: Question bank

Before we start, three words used everywhere below, each earned:

manifold $M$ — the "inside" region you integrate over (an interval, a disk, a solid ball).
boundary $\partial M$ — the "edge" of that region (endpoints, the bounding circle, the bounding surface).
form $ω$ — the thing you integrate; $d ω$ is its exterior derivative (the "local accumulation" — slope, curl, or divergence depending on dimension).

True or false — justify

A boundary can itself have a boundary.

False.

\partial (\partial M) = \emptyset

always — the edge of an edge is empty (a disk's boundary is a circle, and a circle has no endpoints). This is the geometric partner of

d^{2} = 0

(see d squared equals zero).

Generalized Stokes needs the region

M

to be flat.

False. It holds on any smooth oriented

k

-manifold — curved surfaces, solid balls, twisted paths. Flatness is only how Green's theorem looks; the identity itself is coordinate-free.

Every classical theorem (FTC, Green, Stokes, Divergence) is a special case of one identity.

True. They differ only in the dimension

k

and which form

ω

you feed in; the sentence

\int_{\partial M} ω = \int_{M} d ω

never changes.

If you reverse the orientation of

\partial M

, the two sides still match.

==True, but only if you flip both sides consistently.== Reversing

\partial M

flips

\int_{\partial M} ω

by a sign; the equation stays true because

M

's orientation induces exactly that boundary orientation (see Orientation and Induced Boundary Orientation). Flipping just one side breaks it.

d x \land d y

and

d y \land d x

are the same 2-form.

False. The wedge is antisymmetric:

d x \land d y = - d y \land d x

. They are negatives — this sign is orientation, the difference between counterclockwise and clockwise area.

The Fundamental Theorem of Calculus is "too simple" to be a case of Stokes.

False. It is the

k = 1

case exactly:

\int_{a}^{b} f^{'} d x = f (b) - f (a)

\int_{M} d ω = \int_{\partial M} ω

with

ω = f

0

-form and

\partial [a, b] = {b} - {a}

(see Fundamental Theorem of Calculus).

d (d ω) = 0

holds only for functions (

0

-forms).

False.

d^{2} = 0

holds for forms of every degree. On

0

-forms it says

\nabla \times \nabla f = 0

; on

1

-forms it says

\nabla \cdot (\nabla \times F) = 0

— same fact, two faces.

You can integrate a

1

-form over a solid

3

-dimensional region.

False. A

k

-form is integrated over a

k

-dimensional thing. A

1

-form matches curves, a

2

-form matches surfaces, a

3

-form matches volumes. Degree must equal dimension.

Why questions

Why must interior walls cancel when we sum a boundary quantity over tiny cells?

Each shared wall is traversed twice — once as "out" of one cell, once as "in" of its neighbor — with opposite induced orientation, so the two contributions are exact negatives and vanish. Only the unshared outer wall survives.

Why does

d

on a

1

-form in 2D produce exactly the 2D curl

Q_{x} - P_{y}

and nothing else?

Expanding

d (P d x + Q d y)

gives four terms;

d x \land d x

and

d y \land d y

die, and antisymmetry merges the two survivors into a single

(Q_{x} - P_{y}) d x \land d y

. The wedge algebra does the collapsing.

Why is orientation not an optional decoration but part of the theorem?

It supplies the signs the equation needs: the minus at

a

in FTC, the outward-pointing normal in Divergence. Choose the wrong orientation and the whole answer flips sign — the geometry alone can't tell you which sign is right.

Why is the exterior derivative

d

built to satisfy

d (f) = \sum \frac{\partial f}{\partial x _{i}} d x_{i}

So that on functions it is the gradient — the "local rate of change." Everything else (

d

on higher forms giving curl and div) is forced by demanding

d

be linear, agree with the gradient, and respect the wedge. See Exterior Derivative.

Why can one operator

d

play the role of gradient, curl, and divergence?

Because grad/curl/div are the same operation "take the derivative and record how orientation changes," just applied to forms of degree

0

1

, and

2

. The vector-calculus names hide that it's one machine.

Why does

\partial^{2} = 0

(boundary of a boundary is empty) mirror

d^{2} = 0

Stokes couples them:

\int_{M} d (d ω) = \int_{\partial M} d ω = \int_{\partial (\partial M)} ω

. If

\partial (\partial M) = \emptyset

the far side is

0

for every

ω

, which forces

d^{2} = 0

— and vice versa. They are dual through the one identity.

Edge cases

What is

\int_{\partial M} ω

when

M

is a closed surface (like a sphere) with no boundary?

It is

0

, because

\partial M = \emptyset

and integrating over nothing gives nothing. Hence

\int_{M} d ω = 0

too — a closed manifold "sees" no net accumulation of any

d ω

What does Stokes say if

ω

is already closed, i.e.

d ω = 0

Then

\int_{\partial M} ω = 0

for every region

M

. The circulation/flux around any boundary vanishes — this is why gradient fields (

ω = df

, so

d ω = d^{2} f = 0

) do zero net work around loops.

What happens to FTC if

a = b

(a degenerate interval)?

Both sides are

0

\int_{a}^{a} f^{'} d x = 0

and

f (a) - f (a) = 0

. The boundary

{a} - {a}

cancels itself — a point counted with both signs. Consistent, just empty.

In the parent's worked example (

F = (x, y)

on the unit disk), why do the two computations give the identical

2 π

Both are the same Stokes identity for the flux form:

\iint_{M} (\nabla \cdot F) d A = \iint_{M} 2 d A = 2 π

(inside reading) equals

\oint_{\partial M} F \cdot \overset{n}{^} d s = \oint 1 d s = 2 π

(boundary reading). One equation, two readings.

F

has a singularity inside

M

(e.g. blows up at a point), can I still apply Stokes directly?

No — the theorem needs

ω

smooth on all of

M

. A singularity means

d ω

isn't defined there; you must excise the bad point (puncture

M

), which adds a new inner boundary and can leave a nonzero residual. This is the seed of index/winding-number phenomena.

Does the identity still hold in dimension 1 million?

Yes — "generalized" means any dimension

k

. Take a

k

-manifold and a

(k - 1)

-form;

\int_{\partial M} ω = \int_{M} d ω

is unchanged. FTC, Green, Stokes, Divergence are just its first four instances.

Connections

Unification — all three theorems as generalized Stokes — the parent this bank tests
Green's Theorem, Stokes' Theorem (curl), Divergence Theorem, Fundamental Theorem of Calculus — the four costumes
Differential Forms and the Wedge Product and Exterior Derivative — the machinery
d squared equals zero — the deepest trap source
Orientation and Induced Boundary Orientation — where the signs live

Question bank — Unification — all three theorems as generalized Stokes

True or false — justify

Spot the error

Why questions

Edge cases

Connections