32. Adv: one theorem in four disguises — a look at differential forms

Four theorems in this course share an unmistakable family resemblance: the Fundamental Theorem of Calculus (§4.3), the Fundamental Theorem of Line Integrals (§12.3), Green's/Stokes' theorem (§12.4, §12.8), and the divergence theorem (§12.9). Every one of them says, in some form, "a derivative summed throughout an interior equals a plain value on the boundary." This closing lesson makes that resemblance exact: all four are literally the same single theorem, written in four different notations, once the right unifying language — differential forms — is in place.

The pattern, laid side by side

\int_a^bf'(x)\,dx=f(b)-f(a)\qquad\text{(FTC, §4.3)}

\int_C\nabla f\cdot d\vec r=f(B)-f(A)\qquad\text{(FTLI, §12.3)}

\oint_CP\,dx+Q\,dy=\iint_D(Q_x-P_y)\,dA\qquad\text{(Green's, §12.4)}

\oint_C\vec F\cdot d\vec r=\iint_S(\text{curl}\,\vec F)\cdot d\vec S\qquad\text{(Stokes', §12.8)}

\iint_S\vec F\cdot d\vec S=\iiint_E(\text{div}\,\vec F)\,dV\qquad\text{(Divergence, §12.9)}

Every left side is an integral of something over a boundary. Every right side is an integral of a derivative of that same something, over the interior that boundary encloses. Differential forms give a single object, \omega, and a single derivative operator, d, for which every line above becomes one identical equation:

\boxed{\int_{\partial\Omega}\omega=\int_\Omega d\omega}

the generalized Stokes' theorem — where \partial\Omega denotes the boundary of a region \Omega.

What a differential form is

A k-form is built from k of the basis symbols dx,dy,dz, wedged together, with a coefficient function in front. This lesson only needs the low-order cases:

• A 0-form is just a scalar function, f.
• A 1-form is \omega=P\,dx+Q\,dy+R\,dz — exactly the integrand already used in every line integral since §12.2.
• A 2-form is \omega=P\,dy\wedge dz+Q\,dz\wedge dx+R\,dx\wedge dy — exactly the integrand pattern behind every flux integral in §12.7.
• A 3-form is f\,dx\wedge dy\wedge dz — an ordinary triple-integral integrand.

The exterior derivative d reproduces gradient, curl, and divergence

d takes a k-form to a (k+1)-form. Applied to each case above:

d of a 0-form (f) gives the 1-form

df=f_x\,dx+f_y\,dy+f_z\,dz

— exactly the gradient, §10.5's \nabla f, repackaged as a 1-form.

d of a 1-form (P\,dx+Q\,dy+R\,dz) gives the 2-form

d\omega=(R_y-Q_z)\,dy\wedge dz+(P_z-R_x)\,dz\wedge dx+(Q_x-P_y)\,dx\wedge dy

— exactly curl, §12.5's \nabla\times\vec F, with its three components now sitting as the coefficients of a 2-form instead of a vector's entries.

d of a 2-form gives a 3-form whose single coefficient is P_x+Q_y+R_z — exactly divergence, §12.5's \nabla\cdot\vec F.

With this dictionary, the five theorems above are all literally \int_{\partial\Omega}\omega=\int_\Omega d\omega: FTC and FTLI use \omega=f (a 0-form) with d\omega=df; Green's and Stokes' use \omega=P\,dx+Q\,dy(+R\,dz) (a 1-form) with d\omega giving curl; the divergence theorem uses \omega as a 2-form with d\omega giving divergence. Gradient, curl, and divergence — three operators that looked unrelated in §10.5 and §12.5 — are a single operator d, applied at three different levels.

The punchline: d^2=0

Apply d twice to anything, and the result is always exactly zero. This single fact, stated once in the language of forms, is the source of two identities already proven separately, by separate arguments, in this course:

d(df)=0\quad\Longleftrightarrow\quad\text{curl}(\nabla f)=\vec0

— which is exactly §10.2's Clairaut's theorem (f_{xy}=f_{yx}), since expanding curl of a gradient produces nothing but differences of mixed partials that Clairaut forces to cancel.

d(d\omega)=0\quad\Longleftrightarrow\quad\text{div}(\text{curl}\,\vec F)=0

— exactly §12.5's structural identity, proven there by the same Clairaut cancellation, one level up. Both "coincidences" from earlier in this module are the identical algebraic fact, d^2=0, appearing at two different form-levels — not two separate pieces of luck.

Doing it in Python

Confirming d(df)=0 (curl of a gradient vanishes) directly, for a specific function:

import sympy as sp

x, y, z = sp.symbols('x y z')
f = x**2*y + y*z

grad_f = sp.Matrix([sp.diff(f, x), sp.diff(f, y), sp.diff(f, z)])
print(f"df (as a gradient) = {grad_f.T}")

P, Q, R = grad_f
curl_of_grad = sp.Matrix([
    sp.diff(R, y) - sp.diff(Q, z),
    sp.diff(P, z) - sp.diff(R, x),
    sp.diff(Q, x) - sp.diff(P, y)
])
print(f"d(df) as curl(grad f) = {curl_of_grad.T}")
print("zero everywhere -- d^2 = 0, the same fact as Clairaut's theorem")

Confirming the correspondence table directly: computing gradient, curl, and divergence as ordinary vector calculus, and observing they're the d operator at three consecutive levels:

import sympy as sp

x, y, z = sp.symbols('x y z')
f = x*y*z
F = sp.Matrix([x*y, y*z, z*x])

grad_f = sp.Matrix([sp.diff(f, v) for v in (x, y, z)])
curl_F = sp.Matrix([
    sp.diff(F[2], y) - sp.diff(F[1], z),
    sp.diff(F[0], z) - sp.diff(F[2], x),
    sp.diff(F[1], x) - sp.diff(F[0], y)
])
div_F = sum(sp.diff(F[i], v) for i, v in enumerate((x, y, z)))

print(f"d(0-form f)  = grad f = {grad_f.T}")
print(f"d(1-form F)  = curl F = {curl_F.T}")
print(f"d(2-form F)  = div F  = {div_F}")

Worked example

Confirm that d(d\omega)=0 for \omega built from \vec F(x,y,z)=\langle x^2y,y^2z,z^2x\rangle — the same field §12.5 used to demonstrate \text{div}(\text{curl}\,\vec F)=0.

This is exactly §12.5's worked example, relabeled: \omega=P\,dx+Q\,dy+R\,dz with P=x^2y, Q=y^2z, R=z^2x.

d\omega=\text{curl}\,\vec F=\langle-y^2,-z^2,-x^2\rangle

(computed in §12.5's worked example, unchanged)

d(d\omega)=\text{div}\langle-y^2,-z^2,-x^2\rangle=\frac\partial{\partial x}(-y^2)+\frac\partial{\partial y}(-z^2)+\frac\partial{\partial z}(-x^2)=0+0+0

\boxed{d(d\omega)=0}

Sanity check. This is the identical computation §12.5 already performed — nothing new was computed here at all. That's precisely the lesson's point: what looked like two separate facts (Clairaut's theorem in §10.2, and the div-of-curl identity in §12.5) are the same one-line fact, d^2=0, viewed through the unifying lens of differential forms. ✓

Your turn

1. State which of the five boundary-equals-interior-of-derivative theorems (FTC, FTLI, Green's, Stokes', divergence) corresponds to \omega being a 0-form.

2. Explain, using the d^2=0 principle, why the flux of any curl field through any closed surface must always be zero (this was demonstrated computationally at the end of §12.9 — here, explain it from the forms perspective instead).

3. True or false: differential forms introduce genuinely new calculus, beyond what gradient, curl, and divergence already computed in Modules 10 and 12.

Solutions

1. Two theorems use a 0-form: the Fundamental Theorem of Calculus (\omega=f, a function of one variable, \partial\Omega is the two endpoints of an interval) and the Fundamental Theorem of Line Integrals (\omega=f, a function of several variables, \partial\Omega is the two endpoints of a curve) — both are $\int_{\partial\Omega}\omega=\int_\Omega d\omega$ with \omega a 0-form, differing only in how many variables f depends on and what kind of "interior" (interval versus curve) is being integrated over.

2. A curl field is, by definition, d\omega for some 1-form \omega (curl is exactly d applied to a 1-form, from the correspondence table). The flux of a curl field through a closed surface S=\partial E (the boundary of some solid E) is, by the divergence theorem, \iiint_E\text{div}(\text{curl}\,\vec F)\,dV=\iiint_Ed(d\omega)\,dV — and d(d\omega)=0 always, by the d^2=0 principle, so the integral of zero over any solid is 0. No computation of a specific field is needed — the vanishing is structural, guaranteed by d^2=0 alone.

3. False. Every computation performed under the "differential forms" label in this lesson — df, d\omega for a 1-form, d\omega for a 2-form — reproduced gradient, curl, and divergence exactly, using the identical partial derivatives already computed in §10.5 and §12.5. Differential forms don't add new calculus; they add a unifying notation that reveals gradient, curl, and divergence as one operator d at three levels, and reveals Clairaut's theorem and the div-curl identity as one fact, d^2=0, at two levels — genuine conceptual compression, not new computational content.

Check yourself in code

For f(x,y,z)=x^2y+yz, compute df (as a gradient) and confirm d(df)=0 (curl of the gradient vanishes).

Print exactly this:

df = [2*x*y, x**2 + z, y]
d(df) = [0, 0, 0]

import sympy as sp

x, y, z = sp.symbols('x y z')
f = x**2*y + y*z

grad_f = sp.Matrix([sp.diff(f, x), sp.diff(f, y), sp.diff(f, z)])
print("df = ...")

P, Q, R = grad_f
curl_of_grad = sp.Matrix([
    sp.diff(R, y) - sp.diff(Q, z),
    sp.diff(P, z) - sp.diff(R, x),
    sp.diff(Q, x) - sp.diff(P, y)
])
print("d(df) = ...")

import sympy as sp

x, y, z = sp.symbols('x y z')
f = x**2*y + y*z

grad_f = sp.Matrix([sp.diff(f, x), sp.diff(f, y), sp.diff(f, z)])
print(f"df = {list(grad_f)}")

P, Q, R = grad_f
curl_of_grad = sp.Matrix([
    sp.diff(R, y) - sp.diff(Q, z),
    sp.diff(P, z) - sp.diff(R, x),
    sp.diff(Q, x) - sp.diff(P, y)
])
print(f"d(df) = {list(curl_of_grad)}")

Differential forms reveal that the Fundamental Theorem of Calculus, the Fundamental Theorem of Line Integrals, Green's theorem, Stokes' theorem, and the divergence theorem are one single statement, \int_{\partial\Omega}\omega=\int_\Omega d\omega, with \omega ranging over 0-forms, 1-forms, and 2-forms — and that gradient, curl, and divergence are one operator d at three consecutive levels. The identity d^2=0 then unifies §10.2's Clairaut's theorem and §12.5's div-of-curl identity into a single algebraic fact, rather than two separate coincidences.

That closes both this module and the multivariable portion of the course — vectors, partial derivatives, multiple integrals, and vector calculus, all building on the single-variable foundation from Modules 0 through 8. Next: differential equations, where the unknown is no longer a number but a function, defined implicitly by a relationship with its own derivatives.