Unification — all three theorems as generalized Stokes: Visual walkthrough

We use exactly two brand-new pieces of notation, and we define them the moment they appear:

$\int$ — the "add up over a region" sign. Think a very long S for "Sum".
$\partial M$ — read "the boundary of $M$ ": the outer edge of a region $M$ . The symbol $\partial$ is a curly d; it means "edge of".

Everything else we build with pictures.

Step 1 — What "add up around an edge" even means

WHAT: we are defining the left-hand side $\int_{\partial M} ω$ in the simplest possible way — a walk around an edge, summing a push.

WHY: before we can claim "edge equals inside", we must know precisely what "add up along the edge" is. It is not area; it is a signed walk.

PICTURE: below, the blue loop is our path. The small coral arrows are the wind $F$ . The green tick shows the direction we walk (counterclockwise, our chosen orientation). Where wind and walk agree we add a positive amount; where they oppose, a negative amount.

Figure — Unification — all three theorems as generalized Stokes

Step 2 — One tiny square, walked around

WHAT: shrink the loop to a single tiny square with corners spaced $Δ x$ across and $Δ y$ tall. Walk its four sides counterclockwise: right along the bottom, up the right, left along the top, down the left.

WHY: if we can understand the circulation of one tiny square, we can tile any region out of tiny squares and add them up. This is the whole strategy.

PICTURE: the four coloured arrows are the four legs of the walk. Read the sign carefully: the bottom leg walks +x, the top leg walks −x (leftward), so the top leg's contribution comes in with a minus.

Let the wind be $F = (P, Q)$ , meaning $P$ is its rightward part and $Q$ its upward part. Add the four legs:

Group the $P$ terms and the $Q$ terms:

[Q (x + Δ x, y) - Q (x, y)] Δ y - [P (x, y + Δ y) - P (x, y)] Δ x .

Step 3 — Why a derivative is forced to appear

WHAT: each bracket is "value here minus value a step away". Divide and multiply by the step size:

\frac{Q ( x + Δ x , y ) - Q ( x , y )}{Δ x} Δ x Δ y - \frac{P ( x , y + Δ y ) - P ( x , y )}{Δ y} Δ y Δ x .

WHY the derivative, and not some other tool? A derivative is exactly the machine that answers "how fast does this quantity change as I take one tiny step in a given direction?" Our brackets literally ask that. So the derivative is not a clever trick — it is the only tool whose definition matches "value here minus value a step away, per unit step". We write $\partial_{x} Q$ for "rate of change of $Q$ as $x$ increases" (the curly $\partial$ again: change with respect to).

As the square shrinks ( $Δ x, Δ y \to 0$ ) each ratio becomes a derivative:

PICTURE: a paddlewheel in the tiny square. The right-hand side edge feels more/less up-wind than the left ( $\partial_{x} Q$ ); the top edge feels more/less right-wind than the bottom ( $- \partial_{y} P$ ). Their difference spins the wheel.

So for one square: walk-around-the-edge $=$ (local spin) × (area inside). That is already the whole theorem in miniature.

Step 4 — Tile the region, and watch the walls cancel

WHAT: cover the whole region $M$ with a grid of these tiny squares. Add the circulation of every square. This equals the sum of (local spin × area) over all squares — the right-hand side.

WHY: we want the outer boundary, not a mess of tiny squares. The magic is that the inside walls delete themselves.

PICTURE: two neighbouring squares share a wall. The left square walks that shared wall upward; the right square walks the same wall downward. Same wind, opposite steps — they add to zero. This happens on every interior wall. Only edges with no neighbour — the true outer boundary $\partial M$ — survive.

Step 5 — Name the pieces: the form $ω$ and its derivative $d ω$

WHAT: package the walk-integrand and the spin into two objects so the equation gets short.

The thing we integrate along the edge is $ω = P d x + Q d y$ , called a 1-form (a "thing you integrate along a curve"). Here $d x$ means "the leftover rightward step" and $d y$ "the upward step" — the same tiny steps from Step 2.
Applying the exterior derivative $d$ (defined next) turns $ω$ into $d ω = (\partial_{x} Q - \partial_{y} P) d x \land d y$ — exactly the spin-times-area from Step 3.

WHY the wedge $\land$ ? Area is signed: a square swept $x$ -then- $y$ has opposite orientation to $y$ -then- $x$ . The wedge builds this in with one rule, $d x \land d y = - d y \land d x$ . Setting the two equal forces $d x \land d x = 0$ : a "square" with zero width has zero area.

PICTURE: the wedge as an oriented parallelogram — swap the two edge-arrows and the shaded area flips its sign (shown by flipping colour).

Now expand $d ω$ and see the four terms of Step 2 reappear and collapse:

d ω = d of the P d x piece (P_{x} d x + P_{y} d y) \land d x + d of the Q d y piece (Q_{x} d x + Q_{y} d y) \land d y .

Kill $d x \land d x = 0$ and $d y \land d y = 0$ ; use $d y \land d x = - d x \land d y$ :

d ω = (Q_{x} - P_{y}) d x \land d y .

The four terms of the tiny-square walk (Step 2) are these four terms. The theorem's two sides are the same computation seen from outside vs inside.

Step 6 — Every dimension is the same picture

WHAT: the exact same "sum around edge = sum of derivative inside" runs in any dimension. Only the props change.

WHY: the cancellation argument of Step 4 never used the number 2. Shared walls always cancel — points, walls, or faces alike.

PICTURE: three stacked panels — a segment (its two endpoints are the boundary), a patch (its loop is the boundary), a solid (its skin is the boundary) — all captioned "edge = derivative inside".

$k$	$M$ is a...	edge $\partial M$	$d$ becomes	classical name
1	segment $[a, b]$	two endpoints ${b} - {a}$	slope $f^{'}$	FTC
2	flat patch	a loop	curl $Q_{x} - P_{y}$	Green
2-in-3D	surface	its rim	curl of $F$	Stokes
3	solid	its skin	divergence $\nabla \cdot F$	Gauss

Step 7 — Degenerate & edge cases (so you never hit a surprise)

The one-picture summary

Everything above compresses into a single image: a region tiled by tiny squares, interior walls greyed-out (cancelled), the surviving outer edge glowing, and the caption $\oint_{\partial M} ω = \iint_{M} d ω$ . The left side is the glowing rim; the right side is the sum of all the little spins inside.

Recall Feynman retelling — the whole walkthrough in plain words

You want to know the total "swirl" the wind is making inside a fenced field. Two honest ways to count it. Way one (inside): chop the field into tiny squares, drop a paddlewheel in each, add up how much every wheel spins. Way two (edge): just walk once around the outer fence and add up how much the wind pushes you along. They give the same number — and here's why: when you walk around every tiny square, the fence between two neighbouring squares gets walked once in each direction, so those inner fences all cancel. The only fence left un-cancelled is the outer one. So "add spin over all the inside" $=$ "walk the outer fence once". That's the whole theorem: edge outside $=$ derivative inside. Now stretch this idea. In 1-D the "field" is a line segment, its "fence" is just its two endpoints, and the spin becomes the slope — that's the Fundamental Theorem of Calculus. In 3-D the "field" is a solid lump, its "fence" is its skin, and the spin becomes divergence (how much stuff is created per unit volume) — that's the Divergence theorem. Same sentence, different-sized rooms. Write it once and for all: $\int_{\partial M} ω = \int_{M} d ω$ .

Connections

Green's Theorem — the $k = 2$ case built here in Steps 2–5
Stokes' Theorem (curl) — the same tiling on a curved surface
Divergence Theorem — Step 6, the $k = 3$ case
Fundamental Theorem of Calculus — Step 6/7, the $k = 1$ case
Differential Forms and the Wedge Product — Step 5
Exterior Derivative and d squared equals zero
Orientation and Induced Boundary Orientation — Steps 1, 4, 7