06.01.10 · riemann-surfaces / complex-analysis

Cauchy-Riemann equations and harmonic conjugate

shipped3 tiersLean: partial

Anchor (Master): d'Alembert 1752; Euler 1757, 1777; Cauchy 1814, 1825; Riemann 1851 dissertation; Wirtinger 1927; Hartogs 1906; Hörmander *An Introduction to Complex Analysis in Several Variables*; Ahlfors §§2.1, 3.2

Intuition [Beginner]

A complex-valued function $f (z)$ of a complex input has both a real part $u$ and an imaginary part $v$ . Write $z = x + i y$ , so $f (z) = u (x, y) + i v (x, y)$ . The function is called holomorphic when it has a complex derivative — a single complex number $f^{'} (z_{0})$ that describes how $f$ stretches and rotates the plane near $z_{0}$ . The Cauchy-Riemann equations are the pair of conditions on $u$ and $v$ that make this possible:

u_{x} = v_{y}, u_{y} = - v_{x} .

Picture a holomorphic function as a local rotate-and-zoom of the plane. The real part $u$ moves left-to-right and the imaginary part $v$ moves up-and-down; for the combined motion to be a single complex multiplication, the way $u$ varies in $x$ must mirror the way $v$ varies in $y$ , and the way $u$ varies in $y$ must be the negative of the way $v$ varies in $x$ . The two equations encode that mirror relationship.

The coupling has a striking consequence. Knowing one of the two real parts up to a constant determines the other. Given a harmonic function $u$ (a real function with a flat average — see the worked example), there is essentially one $v$ that fits together with it to make a holomorphic $f = u + i v$ . The $v$ is called the harmonic conjugate of $u$ .

Visual [Beginner]

A small square in the $z$ -plane is mapped by a holomorphic $f$ to a small square in the $f (z)$ -plane, rotated and scaled but not skewed. A small square mapped by a non-holomorphic function deforms into a parallelogram with angles bent.

Worked example [Beginner]

Take $f (z) = z^{2}$ . Write $z = x + i y$ . Then

f (z) = (x + i y)^{2} = x^{2} + 2 i x y + i^{2} y^{2} = (x^{2} - y^{2}) + i (2 x y) .

So the real and imaginary parts are $u (x, y) = x^{2} - y^{2}$ and $v (x, y) = 2 x y$ . Compute the four partial derivatives:

u_{x} = 2 x, u_{y} = - 2 y, v_{x} = 2 y, v_{y} = 2 x .

Now check the two Cauchy-Riemann equations. The first asks $u_{x} = v_{y}$ , and $2 x = 2 x$ — confirmed. The second asks $u_{y} = - v_{x}$ , and $- 2 y = - (2 y)$ — confirmed. The function $z^{2}$ is holomorphic on the entire plane.

For contrast, take $g (z) = \overset{z}{ˉ}$ , the complex conjugate. Writing $\overset{z}{ˉ} = x - i y$ , the real and imaginary parts are $u = x$ and $v = - y$ . The partial derivatives are $u_{x} = 1$ , $u_{y} = 0$ , $v_{x} = 0$ , $v_{y} = - 1$ . The first Cauchy-Riemann equation asks $u_{x} = v_{y}$ , but $1 \neq = - 1$ . The check fails, so $\overset{z}{ˉ}$ is not holomorphic.

What this tells us: the Cauchy-Riemann equations are a quick test for holomorphicity that needs only the four partial derivatives. Polynomials in $z$ pass; the conjugation map $\overset{z}{ˉ}$ fails the test in the simplest possible way.

Check your understanding [Beginner]

Formal definition [Intermediate+]

Let $Ω \subseteq C$ be open, and let $f : Ω \to C$ be a function. Identify $C$ with $R^{2}$ via $z = x + i y \leftrightarrow (x, y)$ and write $f (x + i y) = u (x, y) + i v (x, y)$ with $u, v : Ω \to R$ .

The complex derivative of $f$ at $z_{0} \in Ω$ is the limit

f^{'} (z_{0}) := h \to 0 lim \frac{f ( z _{0} + h ) - f ( z _{0} )}{h}, h \in C ∖ {0},

provided the limit exists independently of the direction along which $h \to 0$ . When it does, $f$ is called complex-differentiable at $z_{0}$ ; if this holds at every point of $Ω$ , $f$ is holomorphic on $Ω$ .

The Cauchy-Riemann equations at $(x_{0}, y_{0})$ are the pair of identities

u_{x} (x_{0}, y_{0}) = v_{y} (x_{0}, y_{0}), u_{y} (x_{0}, y_{0}) = - v_{x} (x_{0}, y_{0}),

where the subscripts denote partial derivatives. In matrix form, the real Jacobian

D f (x_{0}, y_{0}) = (u_{x} v_{x} u_{y} v_{y})_{(x_{0}, y_{0})}

has the structure

D f = (a b - b a) with a = u_{x}, b = v_{x},

which is precisely the matrix of multiplication by the complex number $a + ib$ on $R^{2} ≅ C$ . The Cauchy-Riemann system is the exact algebraic condition for $D f$ to belong to the subring $C \subset Mat_{2} (R)$ of complex-multiplication matrices. ^{[Ahlfors Ch. 2 §2.1]}

Wirtinger derivatives. Introduce the operators

\partial := \frac{1}{2} (\partial_{x} - i \partial_{y}), \overset{ˉ}{\partial} := \frac{1}{2} (\partial_{x} + i \partial_{y}),

acting on $C^{1}$ functions $f : Ω \to C$ . A direct expansion gives

\overset{ˉ}{\partial} f = \frac{1}{2} ((u_{x} - v_{y}) + i (v_{x} + u_{y})),

so the Cauchy-Riemann equations are equivalent to the single complex equation $\overset{ˉ}{\partial} f = 0$ . Equivalently, $f$ is holomorphic iff $f$ is annihilated by $\overset{ˉ}{\partial}$ , the anti-holomorphic Cauchy-Riemann operator.

Harmonicity. A real-valued $C^{2}$ function $u : Ω \to R$ is harmonic if it satisfies Laplace's equation $Δ u := u_{xx} + u_{y y} = 0$ . If $f = u + i v$ is holomorphic with $u, v$ of class $C^{2}$ , then differentiating $u_{x} = v_{y}$ once more in $x$ gives $u_{xx} = v_{y x}$ , and differentiating $u_{y} = - v_{x}$ in $y$ gives $u_{y y} = - v_{x y}$ , so $u_{xx} + u_{y y} = v_{y x} - v_{x y} = 0$ by equality of mixed partials. The same argument gives $Δ v = 0$ . Thus the real and imaginary parts of a holomorphic function are both harmonic. (The smoothness assumption $u, v \in C^{2}$ is automatic once $f$ is known to be holomorphic, since holomorphic functions are real-analytic.)

Harmonic conjugate. Given a harmonic $u : Ω \to R$ on a simply-connected open $Ω$ , a harmonic conjugate of $u$ is a harmonic $v : Ω \to R$ such that $f := u + i v$ is holomorphic. The harmonic conjugate is unique up to additive constant. ^{[Ahlfors Ch. 3 §3.2]}

Counterexamples to common slips

Existence of partials is not enough. The function $f (z) = e^{- 1/ z^{4}}$ for $z \neq = 0$ and $f (0) = 0$ has $u_{x} (0, 0) = u_{y} (0, 0) = v_{x} (0, 0) = v_{y} (0, 0) = 0$ , so the Cauchy-Riemann equations hold at the origin, yet $f$ is not complex-differentiable at $0$ . The converse direction of the Cauchy-Riemann equivalence requires real-differentiability of $u, v$ at the point, not merely existence of partials. The Looman-Menchoff theorem provides a refined converse that weakens this hypothesis substantially.
The harmonic-conjugate domain hypothesis is genuine. On the punctured plane $Ω = C^{*}$ , the harmonic function $u (x, y) = lo g x^{2} + y^{2}$ admits no single-valued harmonic conjugate: any local choice glues into the multivalued $ar g z$ , whose monodromy around the origin is $2 π$ . Simply-connectedness of $Ω$ is needed for global existence.
Anti-holomorphic functions satisfy a sign-flipped Cauchy-Riemann system. The map $f (z) = \overset{z}{ˉ}$ has $u_{x} = 1, v_{y} = - 1$ and $u_{y} = 0, v_{x} = 0$ : the first Cauchy-Riemann equation fails. The anti-holomorphic Cauchy-Riemann system $u_{x} = - v_{y}, u_{y} = v_{x}$ characterises maps annihilated by $\partial$ rather than $\overset{ˉ}{\partial}$ .

Key theorem with proof [Intermediate+]

Theorem (Cauchy-Riemann equivalence). Let $Ω \subseteq C$ be open, $f : Ω \to C$ a function, and $z_{0} = x_{0} + i y_{0} \in Ω$ . Write $f = u + i v$ . Then $f$ is complex-differentiable at $z_{0}$ if and only if both of the following hold:

$u, v$ are real-differentiable at $(x_{0}, y_{0})$ ,
the Cauchy-Riemann equations $u_{x} = v_{y}$ and $u_{y} = - v_{x}$ hold at $(x_{0}, y_{0})$ .

In that case the complex derivative satisfies $f^{'} (z_{0}) = u_{x} (x_{0}, y_{0}) + i v_{x} (x_{0}, y_{0})$ . ^{[Ahlfors Ch. 2 §2.1]}

Proof. ( $\Rightarrow$ ) Assume $f$ is complex-differentiable at $z_{0}$ with derivative $f^{'} (z_{0}) = α + i β$ . The defining limit

h \to 0 lim \frac{f ( z _{0} + h ) - f ( z _{0} )}{h} = α + i β

holds for every direction of $h \to 0$ . Restrict $h$ to the real axis, $h = t \in R$ . The increment in the input is $(x_{0} + t) - (x_{0}) = t$ in the $x$ -coordinate and $0$ in the $y$ -coordinate, so

\frac{f ( z _{0} + t ) - f ( z _{0} )}{t} = \frac{u ( x _{0} + t , y _{0} ) - u ( x _{0} , y _{0} )}{t} + i \frac{v ( x _{0} + t , y _{0} ) - v ( x _{0} , y _{0} )}{t} .

Taking $t \to 0$ , the existence of the complex limit forces the existence of $u_{x} (x_{0}, y_{0})$ and $v_{x} (x_{0}, y_{0})$ , with

α + i β = u_{x} (x_{0}, y_{0}) + i v_{x} (x_{0}, y_{0}) .

Now restrict $h$ to the imaginary axis, $h = i t$ with $t \in R$ . The increment is $0$ in $x$ and $t$ in $y$ , so

\frac{f ( z _{0} + i t ) - f ( z _{0} )}{i t} = \frac{1}{i} (\frac{u ( x _{0} , y _{0} + t ) - u ( x _{0} , y _{0} )}{t} + i \frac{v ( x _{0} , y _{0} + t ) - v ( x _{0} , y _{0} )}{t}) .

Using $1/ i = - i$ , the limit as $t \to 0$ is

α + i β = - i u_{y} (x_{0}, y_{0}) + v_{y} (x_{0}, y_{0}) = v_{y} (x_{0}, y_{0}) - i u_{y} (x_{0}, y_{0}) .

Equating the two expressions for $α + i β$ , the real and imaginary parts give $α = u_{x} = v_{y}$ and $β = v_{x} = - u_{y}$ . These are the Cauchy-Riemann equations, and the formula $f^{'} (z_{0}) = u_{x} + i v_{x}$ falls out. Real-differentiability of $u, v$ at $(x_{0}, y_{0})$ follows from the existence of all four partials together with the complex-differentiability hypothesis, which gives a stronger linear approximation.

( $\Leftarrow$ ) Assume $u, v$ are real-differentiable at $(x_{0}, y_{0})$ and the Cauchy-Riemann equations hold there. Write the Cauchy-Riemann values $a := u_{x} (x_{0}, y_{0}) = v_{y} (x_{0}, y_{0})$ and $b := v_{x} (x_{0}, y_{0}) = - u_{y} (x_{0}, y_{0})$ . Real-differentiability of $u$ gives

u (x_{0} + s, y_{0} + t) - u (x_{0}, y_{0}) = a \cdot s - b \cdot t + o (s^{2} + t^{2}),

and similarly real-differentiability of $v$ gives

v (x_{0} + s, y_{0} + t) - v (x_{0}, y_{0}) = b \cdot s + a \cdot t + o (s^{2} + t^{2}) .

Combine these into $f = u + i v$ with $h = s + i t$ :

f (z_{0} + h) - f (z_{0}) = (a \cdot s - b \cdot t) + i (b \cdot s + a \cdot t) + o (∣ h ∣) .

The bracketed real-linear expression equals $(a + ib) (s + i t) = (a + ib) h$ , as a direct expansion confirms. Therefore

f (z_{0} + h) - f (z_{0}) = (a + ib) h + o (∣ h ∣),

which is precisely the statement that $f$ is complex-differentiable at $z_{0}$ with $f^{'} (z_{0}) = a + ib = u_{x} + i v_{x}$ . $□$

The forward direction uses only two directional limits — real and imaginary — to deduce the four-fold partial-derivative identity. The backward direction repackages the two real total-differential approximations into the single complex linear approximation, with the Cauchy-Riemann identities supplying the matching coefficients $a$ and $b$ that make the real linear map a complex multiplication.

Bridge. The Cauchy-Riemann equivalence is the structural pivot from which the rest of one-complex-variable analysis unfolds. First, it builds toward the Cauchy integral formula and Cauchy's theorem (06.01.02), whose proofs rest on the closedness of the form $f (z) d z$ — closedness that is exactly the Cauchy-Riemann condition in disguise via $\overset{ˉ}{\partial} f = 0$ . Second, it builds toward the theory of harmonic functions on the disk and Poisson integrals (06.01.11): the real part of a holomorphic function is harmonic, and conversely every harmonic function on a simply-connected domain is the real part of a holomorphic function with explicit Poisson-integral representation. Third, it builds toward conformal mapping: the matrix structure $D f = (a b - b a)$ is rotation-and-scaling, so a holomorphic map with non-zero derivative preserves angles, and the Cauchy-Riemann system is the exact local condition for conformality. Fourth, it builds toward several-complex-variables analysis through the Wirtinger formalism $\overset{ˉ}{\partial} f = 0$ : in $n$ variables one demands $\overset{ˉ}{\partial}_{j} f = 0$ for each $j$ , and the resulting Cauchy-Riemann system has the elliptic-overdetermined character that underlies the entire theory of holomorphic functions on $C^{n}$ and complex manifolds. Putting these four together, the Cauchy-Riemann equations are the foundational reason two-dimensional harmonic problems collapse to one-dimensional complex analysis — the central insight Riemann built his theory of complex functions on.

Exercises [Intermediate+]

Exercise 3 (medium, symbolic).

Find the harmonic conjugate of $u (x, y) = x^{2} - y^{2} + 3 x$ on the plane, then identify the resulting holomorphic function $f = u + i v$ in closed form in $z$ .

Hint

Use the Cauchy-Riemann equations as a pair of differential equations for $v$ . Integrate $v_{y} = u_{x}$ in $y$ , then differentiate in $x$ and match against $- u_{y}$ to pin down the constant.

Answer

Compute $u_{x} = 2 x + 3$ and $u_{y} = - 2 y$ . The first Cauchy-Riemann equation $v_{y} = u_{x} = 2 x + 3$ integrates in $y$ to $v (x, y) = 2 x y + 3 y + φ (x)$ for some function $φ$ of $x$ alone. Differentiate in $x$ : $v_{x} = 2 y + φ^{'} (x)$ . The second Cauchy-Riemann equation $v_{x} = - u_{y} = 2 y$ forces $φ^{'} (x) = 0$ , so $φ$ is a constant $c$ . The harmonic conjugate is $v (x, y) = 2 x y + 3 y + c$ . Closed form: $f (z) = u + i v = (x^{2} - y^{2} + 3 x) + i (2 x y + 3 y + c) = z^{2} + 3 z + i c$ . The arbitrary additive constant in $v$ corresponds to the arbitrary imaginary additive constant in $f$ . $□$

Exercise 6 (hard, symbolic).

Prove that on a simply-connected open set $Ω \subseteq C$ , every harmonic $u : Ω \to R$ admits a harmonic conjugate $v$ , unique up to additive constant.

Hint

Consider the differential $1$ -form $ω = - u_{y} d x + u_{x} d y$ on $Ω$ . Compute $d ω$ and use simple-connectedness to find a primitive.

Answer

Define $ω := - u_{y} d x + u_{x} d y$ on $Ω$ . The exterior derivative is $d ω = (- u_{y y} - u_{xx}) d x \land d y = - Δ u d x \land d y = 0$ since $u$ is harmonic. So $ω$ is a closed $1$ -form on the simply-connected $Ω$ . By the Poincaré lemma for simply-connected domains in $R^{2}$ , $ω$ has a primitive: a real $C^{2}$ function $v : Ω \to R$ with $d v = ω$ , i.e. $v_{x} = - u_{y}$ and $v_{y} = u_{x}$ . These are exactly the Cauchy-Riemann equations; setting $f := u + i v$ then satisfies the equivalence theorem's hypotheses and is holomorphic. Uniqueness: if $v_{1}, v_{2}$ are two harmonic conjugates, then $v_{1} - v_{2}$ has $(v_{1} - v_{2})_{x} = (v_{1} - v_{2})_{y} = 0$ , so $v_{1} - v_{2}$ is a constant on the connected $Ω$ . $□$

Exercise 8 (hard, symbolic).

Exhibit a harmonic function on the punctured plane $C^{*} := C ∖ {0}$ that admits no globally single-valued harmonic conjugate, and explain how the obstruction is measured.

Hint

Take $u (x, y) = \frac{1}{2} lo g (x^{2} + y^{2}) = lo g ∣ z ∣$ . Try to integrate the candidate $v$ around a circle enclosing the origin.

Answer

Let $u (x, y) = \frac{1}{2} lo g (x^{2} + y^{2})$ . Compute $u_{x} = x / (x^{2} + y^{2})$ , $u_{y} = y / (x^{2} + y^{2})$ , and verify $Δ u = 0$ on $C^{*}$ — a direct computation gives $u_{xx} + u_{y y} = (y^{2} - x^{2}) / (x^{2} + y^{2})^{2} + (x^{2} - y^{2}) / (x^{2} + y^{2})^{2} = 0$ . The closed $1$ -form $ω = - u_{y} d x + u_{x} d y = (- y d x + x d y) / (x^{2} + y^{2})$ is the angle form $d θ$ , whose integral around the unit circle is $2 π \neq = 0$ . Hence $ω$ has no global primitive on $C^{*}$ . Equivalently, the holomorphic candidate $f = u + i v$ would have to be a branch of $lo g z$ , which has monodromy $2 π i$ around the origin and no single-valued representative on $C^{*}$ . The obstruction is measured by the cohomology class $[ω] \in H_{dR}^{1} (C^{*}; R) ≅ R$ , with value $2 π$ on the standard generator. Simple-connectedness of the domain kills this cohomology and restores existence. $□$

Lean formalization [Intermediate+]

lean_status: partial. The Lean companion file Codex.RiemannSurfaces.ComplexAnalysis.CauchyRiemann records the curriculum-facing statements CauchyRiemann_equivalence_placeholder and harmonicConjugate_exists_placeholder as schematic theorems compiled to True, plus a worked compiling example using Mathlib's DifferentiableAt ℂ API. Mathlib provides Complex.differentiableAt_iff_hasDerivAt, the real-Fréchet restriction DifferentiableAt.restrictScalars, and the underlying real partial-derivative infrastructure; the curriculum-facing Cauchy-Riemann equivalence — that $f = u + i v$ is complex-differentiable iff the Cauchy-Riemann system holds and $u, v$ are real-differentiable, with $f^{'} (z_{0}) = u_{x} + i v_{x}$ — and the harmonic-conjugate existence theorem on a simply-connected domain are not packaged as standalone Mathlib results. See the lean_mathlib_gap field in this unit's frontmatter for the precise contribution roadmap.

[object Promise]

Advanced results [Master]

Looman-Menchoff theorem. The hypothesis that $u, v$ are real-differentiable in the Cauchy-Riemann equivalence theorem can be substantially weakened. The Looman-Menchoff theorem states that if $f : Ω \to C$ is continuous on an open $Ω \subseteq C$ , the four partial derivatives $u_{x}, u_{y}, v_{x}, v_{y}$ exist everywhere on $Ω$ (not necessarily continuous), and the Cauchy-Riemann equations hold pointwise on $Ω$ , then $f$ is holomorphic on $Ω$ . The continuity hypothesis is essential — Menchoff showed it cannot be dropped — and the proof proceeds by a Baire-category argument identifying a large subset of $Ω$ on which $f$ is locally analytic, followed by an analytic-continuation extension. ^{[Ahlfors Ch. 2 §2.1 remarks]}

The $\overset{ˉ}{\partial}$ operator on $C^{\infty}$ and currents. The Wirtinger characterisation $\overset{ˉ}{\partial} f = 0$ extends past $C^{1}$ to distributions: a locally integrable $f : Ω \to C$ is weakly holomorphic if $\overset{ˉ}{\partial} f = 0$ in the sense of distributions, i.e. $\int_{Ω} f \overset{ˉ}{\partial} φ d z d \overset{z}{ˉ} = 0$ for every test function $φ$ compactly supported in $Ω$ . The Weyl-Lewy lemma (Weyl 1940) shows weakly holomorphic functions coincide with classical holomorphic functions: $\overset{ˉ}{\partial} f = 0$ in distributions implies $f$ is smooth and classically holomorphic. This is the prototype of elliptic regularity for systems of overdetermined PDE.

Hartogs's theorem. In several complex variables, holomorphicity is defined as the simultaneous Cauchy-Riemann conditions $\overset{ˉ}{\partial}_{j} f = 0$ for each variable $z_{j}$ . Hartogs (1906) discovered the striking phenomenon that separate holomorphicity — holomorphicity in each variable with the others held fixed — already implies joint holomorphicity (and in particular joint continuity), with no a priori continuity hypothesis. The proof uses subharmonic-function methods and the maximum principle on bidiscs, and the result distinguishes complex analysis sharply from real analysis: a function of two real variables that is separately $C^{\infty}$ in each variable can fail to be jointly continuous. ^{[Hartogs 1906]}

Wirtinger calculus and the complex de Rham complex. The decomposition of the complexified exterior algebra $Λ^{*} T_{C}^{*} Ω$ into types splits the exterior derivative as $d = \partial + \overset{ˉ}{\partial}$ on a complex manifold $Ω$ , with

\partial : A^{p, q} (Ω) \to A^{p + 1, q} (Ω), \overset{ˉ}{\partial} : A^{p, q} (Ω) \to A^{p, q + 1} (Ω) .

The Dolbeault cohomology groups $H_{\overset{ˉ}{\partial}}^{p, q} (Ω)$ measure obstructions to solving $\overset{ˉ}{\partial} u = f$ on $Ω$ . For a smooth $\overset{ˉ}{\partial}$ -closed $(0, 1)$ -form $f$ on $C^{n}$ , the Dolbeault lemma constructs a smooth $u$ with $\overset{ˉ}{\partial} u = f$ on any pseudoconvex open set, generalising the harmonic-conjugate construction beyond simply-connected planar domains.

Hodge theory in the plane. On a compact Riemann surface $X$ (in particular $C_{\infty}$ , see 06.01.07) the operator $\overset{ˉ}{\partial}$ acts on the smooth $(0, 1)$ -forms with Hodge decomposition

A^{0, 1} (X) = H^{0, 1} (X) \oplus im \overset{ˉ}{\partial} \oplus im \overset{ˉ}{\partial}^{*},

with $H^{0, 1} (X)$ the finite-dimensional space of harmonic representatives. On $C_{\infty}$ this group vanishes, recovering the global solvability of $\overset{ˉ}{\partial}$ on the sphere; on a torus $C /Λ$ it is one-dimensional. The Cauchy-Riemann equations sit at the level of $0$ -forms in this complex, with their kernel the holomorphic functions.

Elliptic regularity for general elliptic operators. The Cauchy-Riemann system $\overset{ˉ}{\partial} f = 0$ is the simplest example of an elliptic PDE: the symbol $σ (\overset{ˉ}{\partial}) (ξ) = \frac{1}{2} (ξ_{x} + i ξ_{y})$ is non-zero for every non-zero cotangent vector $ξ \in R^{2} ∖ {0}$ . Solutions of elliptic PDEs with $C^{\infty}$ coefficients are automatically $C^{\infty}$ (interior elliptic regularity), and when the equation has analytic coefficients the solutions are real-analytic (Petrovskii 1939; Morrey 1958). For Cauchy-Riemann this is the statement that holomorphic functions are real-analytic — the bootstrap from $\overset{ˉ}{\partial} f = 0$ to convergent Taylor expansion is the foundational instance of the general elliptic-regularity philosophy. ^{[Hörmander Ch. 1–2]}

Conformality. A real-differentiable map $f : Ω \to R^{2}$ with non-singular Jacobian is conformal at $z_{0}$ (preserves angles and orientation) iff $D f (z_{0})$ is a positive multiple of a rotation matrix — equivalently $D f (z_{0}) = (a b - b a)$ with $a^{2} + b^{2} > 0$ . This is exactly the Cauchy-Riemann condition together with $f^{'} (z_{0}) \neq = 0$ . Hence a holomorphic function with non-vanishing derivative is conformal at every point, and conformality of a smooth orientation-preserving $f$ is equivalent to holomorphicity with $f^{'} \neq = 0$ .

Synthesis. The Cauchy-Riemann equations are the foundational pivot of one-complex-variable analysis, and the synthesis runs in four directions. First, the synthesis with integration theory: the closedness of the form $f (z) d z$ — Cauchy's theorem — is the Cauchy-Riemann equation $\overset{ˉ}{\partial} f = 0$ rephrased through the language of differential forms, and every theorem about complex contour integration is a corollary statement about a closed $1$ -form on a planar domain. Second, the synthesis with harmonic-function theory: the real and imaginary parts of a holomorphic function are harmonic, and on simply-connected domains every harmonic function is the real part of a holomorphic one. Two-dimensional Dirichlet problems for $Δ u = 0$ thereby collapse into one-dimensional problems about holomorphic functions — the structural reason for the disproportionate analytic power of complex analysis in two real dimensions.

Third, the synthesis with conformal mapping: the Cauchy-Riemann condition is precisely the local condition for an orientation-preserving smooth map to preserve angles, and the Riemann mapping theorem 06.01.06 is the global existence statement that simply-connected proper subdomains of $C$ are conformally equivalent to the disk. The Cauchy-Riemann system supplies the differential equation; the Riemann mapping theorem supplies its existence theorem. Fourth, the synthesis with several-complex-variables analysis and complex geometry: $\overset{ˉ}{\partial}$ extends to the Dolbeault complex on $C^{n}$ and on complex manifolds, with the cohomology $H_{\overset{ˉ}{\partial}}^{p, q}$ measuring obstructions to higher-dimensional Cauchy-Riemann problems. Hartogs's separately-holomorphic theorem and Hörmander's $L^{2}$ -estimates for $\overset{ˉ}{\partial}$ are the higher-dimensional descendants of the planar Cauchy-Riemann story, and the elliptic-overdetermined nature of $\overset{ˉ}{\partial}$ in $n$ variables is the analytic backbone of Stein manifolds, sheaf cohomology of coherent analytic sheaves, and Kodaira's vanishing theorems.

Full proof set [Master]

The advanced results assembled above follow from the proofs given in earlier sections together with classical results recorded in Ahlfors Complex Analysis Ch. 2 §2.1 and Ch. 3 §3.2, Hörmander An Introduction to Complex Analysis in Several Variables Chs. 1–2, and the original sources. Five reference points: (a) the Looman-Menchoff theorem (Looman 1923; Menchoff 1936) uses a Baire-category decomposition $Ω = ⋃_{n} F_{n}$ where each $F_{n}$ is the closed set on which the four partials are bounded by $n$ , applying Egorov's theorem and the dominated-convergence approach to verify $f$ is locally given by a Cauchy integral on the interior of each $F_{n}$ , then propagating analyticity via the identity theorem. (b) The Weyl-Lewy lemma proceeds by convolving with a smooth approximate identity to get a $C^{\infty}$ family $f_{ε}$ satisfying $\overset{ˉ}{\partial} f_{ε} = 0$ , hence classically holomorphic, and verifying that $f_{ε} \to f$ in $L_{loc}^{1}$ ; the local-uniform-convergence-of-holomorphic-functions theorem then gives $f$ classically holomorphic. (c) Hartogs's theorem (Hartogs 1906) uses subharmonicity of $lo g ∣ f ∣$ in each variable separately, combined with a Hartogs-figure argument that propagates analyticity from a polydisc to a larger one. (d) The Dolbeault lemma proceeds by reducing to the one-variable Cauchy-Riemann equation $\overset{ˉ}{\partial} u = f$ for $f \in A_{c}^{0, 1} (C)$ compactly supported, where the explicit solution is

u (z) = \frac{1}{2 π i} \int_{C} \frac{f ( ζ ) d ζ \land d ζ ˉ}{ζ - z},

via the Cauchy-Pompeiu formula; iterating over variables and using a partition-of-unity argument gives the multi-variable statement on pseudoconvex domains. (e) Elliptic regularity for $\overset{ˉ}{\partial}$ follows from the Calderón-Zygmund $L^{p}$ -theory of singular integrals applied to the convolution kernel $1/ z$ , combined with a bootstrap from $L^{p}$ to $C^{k, α}$ via the Morrey embedding. ^{[Ahlfors Ch. 2 §2.1; Hörmander Chs. 1–2]}

Connections [Master]

Holomorphic function 06.01.01 — The Cauchy-Riemann equivalence is the precise differential characterisation of holomorphicity: a function is holomorphic iff its real and imaginary parts are real-differentiable and satisfy the Cauchy-Riemann system. Every theorem about holomorphic functions ultimately routes through the Cauchy-Riemann conditions at the level of the local differential geometry, with the equivalence theorem supplying the bridge between the analytic limit-definition and the partial-differential-equation reformulation.
Cauchy integral formula 06.01.02 — Cauchy's theorem $\oint_{γ} f (z) d z = 0$ for a closed contour bounding a region of holomorphicity is the integrated form of the closed- $1$ -form statement $d (f d z) = 0$ , which holds because $\overset{ˉ}{\partial} f = 0$ . The Cauchy integral formula's representation of $f$ at interior points by boundary values is the dual integration-by-parts statement that the operator $\overset{ˉ}{\partial}$ has a fundamental solution $1/ (π z)$ , and the Cauchy-Pompeiu formula extends the representation to non-holomorphic functions with a correction term involving $\overset{ˉ}{\partial} f$ .
Riemann mapping theorem 06.01.06 — The Cauchy-Riemann equations characterise conformality at each point, and the Riemann mapping theorem is the global existence statement that every simply-connected proper subdomain of $C$ is conformally equivalent to the unit disk. The local differential-equation system $u_{x} = v_{y}, u_{y} = - v_{x}$ supplies the pointwise condition; the Riemann mapping theorem supplies the global existence of a function satisfying it that also realises a prescribed conformal equivalence.
Riemann sphere 06.01.07 — On the Riemann sphere $C_{\infty} = P^{1} (C)$ the Cauchy-Riemann equations apply chart-by-chart: a function $C_{\infty} \to C_{\infty}$ is holomorphic iff its expression in each affine chart satisfies the Cauchy-Riemann system. The transition map $z \mapsto 1/ z$ is itself a Möbius transformation, hence holomorphic, and the chart-compatibility of the Cauchy-Riemann conditions is what allows the local differential characterisation to globalise into the Riemann-surface structure.
Harmonic functions on the disk and Poisson integrals [06.01.11, pending] — The real and imaginary parts of a holomorphic function are harmonic, and the harmonic-conjugate construction inverts this: given a harmonic $u$ on a simply-connected domain, there is a holomorphic $f$ with $Re f = u$ unique up to imaginary additive constant. The Poisson integral on the disk extends this to harmonic functions with prescribed boundary values, and the conjugate Poisson kernel produces the harmonic conjugate $v$ from the boundary data of $u$ — a constructive solution of the harmonic-conjugate problem.

Historical & philosophical context [Master]

Jean le Rond d'Alembert wrote down the equations now bearing Cauchy and Riemann's names in his Essai d'une nouvelle théorie de la résistance des fluides (1752), as the conditions for a planar fluid flow to be both incompressible and irrotational — the velocity field $(u, - v)$ then has potential $u$ and stream function $v$ with the relations $u_{x} = v_{y}$ and $u_{y} = - v_{x}$ . Leonhard Euler, in Principia motus fluidorum (1757) and De repraesentatione superficiei sphaericae super plano (1777), gave the equations a complex-function-theoretic reading, observing that the conditions characterise functions of $x + i y$ in a sense that he made precise for analytic functions defined by power series. ^{[Cauchy 1825]}

Augustin-Louis Cauchy placed complex differentiation on systematic ground in his Mémoire sur les intégrales définies (1814) and the foundational Mémoire sur les intégrales définies prises entre des limites imaginaires (1825), where he treated the equations as identities that hold for differentiable complex functions and proved the integral theorem now bearing his name. Bernhard Riemann, in his 1851 Inaugural Dissertation Grundlagen für eine allgemeine Theorie der Functionen einer veränderlichen complexen Grösse at Göttingen under Gauss, made the Cauchy-Riemann equations the defining property of complex-analytic — he called them monogenic — functions, breaking from the power-series approach of Cauchy and Weierstrass in favour of a geometric-function-theoretic standpoint. Riemann's choice of starting point is the structural origin of the modern view that holomorphicity is a local differential-equation property of a complex function, rather than a convergent-series property. ^{[Riemann 1851 dissertation]}

Wilhelm Wirtinger, in Zur formalen Theorie der Funktionen von mehr komplexen Veränderlichen (1927), introduced the operators $\partial = \frac{1}{2} (\partial_{x} - i \partial_{y})$ and $\overset{ˉ}{\partial} = \frac{1}{2} (\partial_{x} + i \partial_{y})$ , condensing the Cauchy-Riemann system to the single complex equation $\overset{ˉ}{\partial} f = 0$ and opening the modern several-complex-variables and complex-manifold formulations. Friedrich Hartogs's discovery (1906) that separate holomorphicity implies joint holomorphicity initiated the systematic study of holomorphic functions in several variables, and Lars Hörmander's $L^{2}$ -estimates for $\overset{ˉ}{\partial}$ in the 1960s completed the analytic-elliptic-PDE programme that the Cauchy-Riemann equations seeded two centuries earlier.

Bibliography [Master]

d'Alembert, J. le R., Essai d'une nouvelle théorie de la résistance des fluides, Paris (1752).
Euler, L., Principia motus fluidorum, Novi Comm. Acad. Sci. Petrop. 6 (1761), 271–311 [presented 1757].
Euler, L., De repraesentatione superficiei sphaericae super plano, Acta Acad. Sci. Petrop. 1 (1777), 107–132.
Cauchy, A.-L., Mémoire sur les intégrales définies, Mém. présentés par divers savants Acad. roy. sci. Inst. France 1 (1827, presented 1814).
Cauchy, A.-L., Mémoire sur les intégrales définies prises entre des limites imaginaires, Paris (1825).
Riemann, B., Grundlagen für eine allgemeine Theorie der Functionen einer veränderlichen complexen Grösse (Inaugural Dissertation), Göttingen (1851).
Hartogs, F., Zur Theorie der analytischen Funktionen mehrerer unabhängiger Veränderlichen, Math. Ann. 62 (1906), 1–88.
Wirtinger, W., Zur formalen Theorie der Funktionen von mehr komplexen Veränderlichen, Math. Ann. 97 (1927), 357–375.
Ahlfors, L. V., Complex Analysis, 3rd ed., McGraw-Hill (1979). Ch. 2 §2.1; Ch. 3 §3.2.
Hörmander, L., An Introduction to Complex Analysis in Several Variables, 3rd ed., North-Holland (1990).
Conway, J. B., Functions of One Complex Variable I, 2nd ed., GTM 11, Springer (1978).
Stein, E. M. and Shakarchi, R., Complex Analysis, Princeton Lectures in Analysis II, Princeton (2003).
Needham, T., Visual Complex Analysis, Oxford University Press (1997). Ch. 4.

Prerequisites

06.01.07

Tier anchors

beginner: Needham *Visual Complex Analysis* §4.II (amplitwist); 3Blue1Brown 'derivative of a complex function'
intermediate: Ahlfors *Complex Analysis* Ch. 2 §2.1 and Ch. 3 §3.2; Stein-Shakarchi Vol. II §1.2; Conway *Functions of One Complex Variable* §III.2
master: d'Alembert 1752; Euler 1757, 1777; Cauchy 1814, 1825; Riemann 1851 dissertation; Wirtinger 1927; Hartogs 1906; Hörmander *An Introduction to Complex Analysis in Several Variables*; Ahlfors §§2.1, 3.2

References

TODO_REF
Ahlfors — Complex Analysis · Ch. 2 §2.1 (Cauchy-Riemann equations) and Ch. 3 §3.2 (harmonic conjugate)
TODO_REF
Riemann — Grundlagen für eine allgemeine Theorie der Functionen einer veränderlichen complexen Grösse (1851) · Inaugural Dissertation, Göttingen; CR equations as the defining property of complex (monogenic) functions
TODO_REF
Cauchy — Mémoire sur les intégrales définies prises entre des limites imaginaires (1825) · originating reference for complex differentiation on systematic grounds
TODO_REF
Wirtinger — Zur formalen Theorie der Funktionen von mehr komplexen Veränderlichen (1927) · Math. Ann. 97; the $\partial / \bar\partial$ operator formalism
TODO_REF
Hörmander — An Introduction to Complex Analysis in Several Variables · Ch. 1–2, $\bar\partial$ equation and elliptic-regularity perspective

Lean module

Codex.RiemannSurfaces.ComplexAnalysis.CauchyRiemann

Mathlib gap

Mathlib has complex differentiability via `DifferentiableAt ℂ` together with `Complex.differentiableAt_iff_hasDerivAt`, the real-Fréchet restriction `DifferentiableAt.restrictScalars`, and the underlying real partial-derivative API, but the curriculum-facing Cauchy-Riemann equivalence — that `f = u + i v` is complex-differentiable at $z_0$ iff $u, v$ are real-differentiable at $(x_0, y_0)$ with $u_x = v_y$ and $u_y = -v_x$, with $f'(z_0) = u_x + i v_x$ — is not packaged as a standalone theorem. The harmonic-conjugate existence statement on a simply-connected open set, the Wirtinger $\partial / \bar\partial$ characterisation $\bar\partial f = 0 \Leftrightarrow f$ holomorphic, and Hartogs's separately-holomorphic theorem are likewise absent. Building these would require an explicit `Complex.realPartDeriv` API, a `harmonicConjugate` constructor via line-integration on simply-connected domains, and the Wirtinger derivative operators as Mathlib definitions.

Reviewer

TBD

Estimated time

beginner: 15m
intermediate: 35m
master: 65m