Altanis

1.6Complex-Valued Functions as Mappings

Updated 2 Aug 2026Chapter (PDF)

[1.6.1]Remark(Real Differentiability vs. Complex Differentiability)#

We now make the distinction between treating a complex-valued function as a function R2R2\bR^2 \to \bR^2. Consider the function f(z)=zf(z) = \bar{z}, then note

f(z)=limh0f(z+h)f(z)h=limh0hh,f'(z) = \lim_{h \to 0} \frac{f(z + h) - f(z)}{h} = \lim_{h \to 0} \frac{\bar{h}}{h},

which does not exist (parameterizing z/z\bar{z}/z by an angle yields exp(2iθ)\exp(-2i\theta), and so the limit varies based on path chosen). Thus f:CCf: \bC \to \bC is not holomorphic. But now consider g:R2R2g: \bR^2 \to \bR^2 defined by g(x,y)=(u(x,y),v(x,y))=(x,y)g(x, y) = (u(x, y), v(x, y)) = (x, -y). This function, as a real-valued one, is linear, infinitely differentiable, and enjoys many properties. One such property is that its total derivative (as a linear map), expressed in the standard basis of R2\bR^2, is written as

g=[uxuyvxvy]=[1001].g' = \begin{bmatrix} \frac{\partial u}{\partial x} & \frac{\partial u}{\partial y} \\ \frac{\partial v}{\partial x} & \frac{\partial v}{\partial y} \end{bmatrix} = \begin{bmatrix} 1 & 0 \\ 0 & -1 \end{bmatrix}.

From the perspective of multivariate theory, a map g:R2R2g: \bR^2 \to \bR^2 has a derivative that is a linear map in two-dimensions, in contrast with a map f:CCf: \bC \to \bC having a singular, complex number for its derivative (that is also a one-dimensional linear map).

[1.6.2]Remark(Cauchy-Riemann Equations)#

We can link real differentiability and complex differentiability, however. For a complex function f:ΩCf: \Omega \to \bC to be differentiable, its difference function must take on the same value as h0h \to 0 through any path. Writing h=h1+ih2h = h_1 + ih_2 and z=x+iyz = x + iy (and so f(z)=f(x,y)f(z) = f(x, y)), we may choose a point z0=x0+iy0Ωz_0 = x_0 + i y_0 \in \Omega for which f(z0)f'(z_0) exists and yield the following:

f(z0)=limh0f(z0+h)f(z0)h=limh10f(x0+h1,y0)f(x0,y0)h1=fx(x0)(h2=0)=limh20f(x0,y0+h2)f(x0,y0)ih2=1ify(x0)(h1=0).\begin{align*} f'(z_0) = \lim_{h \to 0} \frac{f(z_0 + h) - f(z_0)}{h} &= \lim_{h_1 \to 0} \frac{f(x_0 + h_1, y_0) - f(x_0, y_0)}{h_1} = {\frac{\partial f}{\partial x}} (x_0) \quad (h_2 = 0) \\ &= \lim_{h_2 \to 0} \frac{f(x_0, y_0 + h_2) - f(x_0, y_0)}{ih_2} = {\frac{1}{i}\frac{\partial f}{\partial y}} (x_0) \quad (h_1 = 0). \\ \end{align*}

If ff is holomorphic, we have that fx=1ify=ify\frac{\partial f}{\partial x} = \frac{1}{i} \frac{\partial f}{\partial y} = -i\frac{\partial f}{\partial y} by the uniqueness of the derivative. If we write f(z)=u(z)+iv(z)f(z) = u(z) + iv(z), then we have that the partials ux,uy,vx,vyu_x, u_y, v_x, v_y all exist, and they are such that

fx=ify    (ux)+i(vx)=i[(uy)+i(vy)]=(vy)+i(uy),\frac{\partial f}{\partial x} = -i\frac{\partial f}{\partial y} \implies \left(\frac{\partial u}{\partial x}\right) + i\left(\frac{\partial v}{\partial x}\right) = i\left[ \left(\frac{\partial u}{\partial y}\right) + i \left(\frac{\partial v}{\partial y}\right) \right] = \left(-\frac{\partial v}{\partial y}\right) + i\left(\frac{\partial u}{\partial y}\right),

implying the Cauchy-Riemann equations:

ux=vyuy=vx.\frac{\partial u}{\partial x} = \frac{\partial v}{\partial y} \quad \frac{\partial u}{\partial y} = -\frac{\partial v}{\partial x}.
[1.6.3]Remark(Derivation of Wirtinger Operators)#

We will motivate the introduction of two important differential operators known as Wirtinger operators. To get started, let zCz \in \bC be a complex variable. Noting that CR2\bC \cong \bR^2, write z=x+iyz = x + iy and z=xiy\bar{z} = x - iy. We may “solve” for x,yx, y in terms of z,zz, \bar{z} as such:

x=z+z2y=zz2i.x = \frac{z + \bar{z}}{2} \quad y = \frac{z - \bar{z}}{2i}.

Treating z\bar{z} as independent of zz, we can expand the differential operators z\frac{\partial}{\partial z} and z\frac{\partial}{\partial \bar{z}} by the multivariate chain rule.

z=xz1/2x+yz1/2iy=12(x+1iy),\frac{\partial}{\partial z} = \underbrace{\frac{\partial x}{\partial z}}_{1/2} \frac{\partial}{\partial x} + \underbrace{\frac{\partial y}{\partial z}}_{1/2i} \frac{\partial}{\partial y} = \frac{1}{2} \left( \frac{\partial}{\partial x} + \frac{1}{i} \frac{\partial}{\partial y} \right),
z=xz1/2x+yz1/2iy=12(x11y).\frac{\partial}{\partial \bar{z}} = \underbrace{\frac{\partial x}{\partial \bar{z}}}_{1/2} \frac{\partial}{\partial x} + \underbrace{\frac{\partial y}{\partial \bar{z}}}_{-1/2i} \frac{\partial}{\partial y} = \frac{1}{2} \left( \frac{\partial}{\partial x} - \frac{1}{1} \frac{\partial}{\partial y} \right).
[1.6.4]Proposition#

If f(z)=u(z)+iv(z)f(z) = u(z) + iv(z) is holomorphic at z0z_0, then

fz(z0)=0f(z0)=fz(z0)=2uz(z0).{\frac{\partial f}{\partial \bar{z}}} (z_0) = 0 \quad f'(z_0) = {\frac{\partial f}{\partial z}} (z_0) = 2 {\frac{\partial u}{\partial z}} (z_0).

Let FF be a mapping in terms of real variables such that F(x,y)=f(x+iy)F(x, y) = f(x + iy). Then we have that FF is differentiable, and its derivative can be represented by a Jacobian JFJ_F such that

det(JF(x0,y0))=f(x0+iy0)2.\det(J_F(x_0, y_0)) = |f(x_0 + i y_0)|^2.
Proof.

Suppose ff is holomorphic at z0z_0. Then ff satisfies the Cauchy-Riemann equations, meaning fx(z0)=1ify(z0){\frac{\partial f}{\partial x}} (z_0) = \frac{1}{i} \frac{\partial f}{\partial y} (z_0), and so

fz(z0)=12(fx(z0)1ify(z0))=0.{\frac{\partial f}{\partial \bar{z}}} (z_0) = \frac{1}{2}\left( {\frac{\partial f}{\partial x}} (z_0) - \frac{1}{i} {\frac{\partial f}{\partial y}} (z_0) \right) = 0.

From this, we have that

f(z0)=12(fx(z0)+fx(z0))=12(fx(z0)+1ify(z0))=fz(z0).f'(z_0) = \frac{1}{2}\left( {\frac{\partial f}{\partial x}}(z_0) + {\frac{\partial f}{\partial x}}(z_0) \right) = \frac{1}{2}\left( {\frac{\partial f}{\partial x}}(z_0) + \frac{1}{i} {\frac{\partial f}{\partial y}}(z_0) \right) = {\frac{\partial f}{\partial z}}(z_0).

From the Cauchy-Riemann Equations, we get that fz=2uz\frac{\partial f}{\partial z} = 2\frac{\partial u}{\partial z} for f=u+ivf = u + iv. We have already proved that complex differentiability implies real differentiability by the Cauchy-Riemann equations. Finally, note that

JF(x0,y0)=[ux(x0)uy(y0)vx(x0)vy(y0)],J_F(x_0, y_0) = \begin{bmatrix} u_x(x_0) & u_y(y_0) \\ v_x(x_0) & v_y(y_0) \end{bmatrix},

and so det(JF(x0,y0))=ux(x0)vy(y0)uy(y0)vx(x0)=(ux(x0))2+(uy(y0))2=f(z0)2\det(J_F(x_0, y_0)) = u_x(x_0) v_y(y_0) - u_y(y_0)v_x(x_0) = (u_x(x_0))^2 + (u_y(y_0))^2 = |f'(z_0)|^2.

[1.6.5]Theorem(Criterion for Holomorphicity)#

A function f(z)=u(z)+iv(z)f(z) = u(z) + iv(z) is holomorphic if and only if u,vu, v have continuous, first-order partial derivatives that satisfy the Cauchy-Riemann equations.

Proof.

():(\Longrightarrow): This is implied from taking the limit of the difference quotient along R\bR and iRi\bR.

():(\Longleftarrow): Suppose u,vu, v have continuous first-order partial derivatives that satisfy the Cauchy-Riemann equations. From multivariate theory, note that

u(x+h,y+k)u(x,y)=uxh+uyk+o1(h+ik),v(x+h,y+k)v(x,y)=vxh+vyk+o2(h+ik).\begin{align*} u(x + h, y + k) - u(x, y) &= u_x h + u_y k + o_1(h + ik),\\ v(x + h, y + k) - v ( x, y) &= v_x h + v_y k + o_2(h + ik). \end{align*}

Then note

f(z+[h+ik])f(z)=(ux+ivx)(h+ik)+[o1(h+ik)+io2(h+ik)],f(z + [h + ik]) - f(z) = (u_x + iv_x)(h + ik) + [o_1(h + ik) + i o_2(h + ik)],

meaning

f(z)=limh+ik0f(z+[h+ik])f(z)h+ik=ux+ivx.f'(z) = \lim_{h + ik \to 0} \frac{f(z + [h + ik]) - f(z)}{h + ik} = u_x + i v_x.