Why a graph fails — and a map saves us
In single-variable calculus you pictured a function as a graph: feed in x on the horizontal axis, read off y on the vertical axis, and trace a curve. A complex function f sends a complex input z to a complex output w = f(z). But z already lives in a two-dimensional plane, and so does w — to graph the pair (z, w) honestly you would need four dimensions, which no one can draw. The graph picture quietly dies here.
So we change our whole way of seeing. Instead of one plane with a curve on it, we draw two planes side by side: a z-plane on the left, holding all the inputs, and a w-plane on the right, holding all the outputs. The function becomes a mapping — an arrow-by-arrow recipe that picks up each point of the z-plane and sets it down somewhere in the w-plane. We stop asking "what is the height above this point?" and start asking "where does this point go?"
Splitting into real and imaginary parts
Write the input as z = x + i y, with x and y its real and imaginary parts. The output w = f(z) is itself a complex number, so it too splits into a real part and an imaginary part. Each of those parts is a plain real-valued function of the two real variables x and y. We name them u and v and write f = u + i v. This single move is the bridge from complex analysis back to the multivariable calculus you already know.
z = x + i y f(z) = u(x, y) + i v(x, y) example: f(z) = z^2 z^2 = (x + i y)^2 = (x^2 - y^2) + i (2 x y) so u(x, y) = x^2 - y^2, v(x, y) = 2 x y
Working out u and v is your reliable first move on any new function — it converts a mysterious complex object into two concrete surfaces you could plot, and later it is exactly what the Cauchy-Riemann equations will be written in terms of. Be honest, though, about what this split does and does not buy you: knowing u and v as separate real functions tells you everything about where points go, but it hides the special rigidity that makes a function complex-differentiable. That deeper structure is the story of the next rung; here we are only learning to read the map.
A zoo of simple maps
The fastest way to build intuition is to watch a handful of small maps act on the plane. Start with f(z) = z + a, where a is a fixed complex number. Every point slides by the same arrow a — the whole plane shifts without rotating or stretching. This is the translation, the gentlest map there is: a circle stays the same circle, just relocated.
Next take f(z) = a z. Writing a = r e^(i theta) in polar form, multiplying by a scales every modulus by r and adds theta to every argument — so the plane spins about the origin by angle theta and dilates by the factor r at once. This rotation-and-scaling is the geometric meaning of complex multiplication you met on the previous rung, now seen as a map. The conjugation map f(z) = z-bar is different in kind: it reflects every point across the real axis, flipping the upper half-plane down onto the lower. Reflection reverses orientation, which will matter a great deal once we care about angles.
Now the two that genuinely bend the plane. The squaring map f(z) = z^2 doubles every argument, so it wraps the plane around itself twice: a wedge of angle pi opens up into a full half-turn, and the two opposite points z and -z land on the same output. And inversion f(z) = 1/z sends modulus |z| to 1/|z| while negating the argument — points near the origin fly far out, points far away crowd in near the origin, and the unit circle stays put. We will spend the whole next guide drawing exactly what z^2 and 1/z do to lines and circles; for now just feel the flavour of each.
The complex limit: every road must agree
To say anything precise about a map we need a notion of nearness, and that means a limit. We write that f(z) approaches L as z approaches z_0, meaning the limit of a complex function equals L, when we can force the distance |f(z) - L| as small as we like by taking z close enough to z_0 — measured by the modulus |z - z_0|, which is just ordinary distance in the plane. The definition reads exactly like the real one, with modulus in place of absolute value.
But there is a quiet, crucial difference. On the real line you could only approach z_0 from two sides, left and right. In the plane you can approach z_0 along infinitely many directions — straight in from the east, in from the north, spiralling, wiggling. For the limit to exist, every one of these approaches must give the same answer L. This is the directional independence that makes complex limits genuinely more demanding than real ones, and it is the seed of everything strong and rigid that complex analysis will later deliver.
Here is the test in action. Consider g(z) = z-bar / z away from the origin, and ask whether it has a limit as z approaches 0. Come in along the positive real axis, where z = x is real: then z-bar = x too, so g = x/x = 1. Now come in along the imaginary axis, where z = i y: then z-bar = -i y, so g = (-i y)/(i y) = -1. Two roads, two different answers — so no limit exists at 0. A real-variable eye would have missed this entirely; the plane's extra freedom of approach is exactly what exposes it.
Continuity, and the curves a map carries
Once limits are in hand, continuity reads just as you would hope: f is continuous at z_0 when the limit of f(z) as z approaches z_0 exists and equals f(z_0) — no jumps, no surprises. The picture is the one from any calculus course: nearby inputs go to nearby outputs, so the map never tears the plane.
Because everything splits as f = u + i v, complex continuity is exactly equivalent to the joint continuity of the two real functions u(x, y) and v(x, y). That is genuinely useful: it lets you import every continuity fact from multivariable calculus wholesale. Sums, products, and compositions of continuous complex functions are continuous, and a quotient is continuous wherever the denominator is non-zero — so polynomials are continuous everywhere, and 1/z is continuous everywhere except at the origin, precisely where it blows up.
Finally, the objects a map most vividly acts on are not single points but curves. Describe a path by letting a real parameter t run over an interval and writing z(t) = x(t) + i y(t); as t sweeps along, z(t) traces a road through the z-plane. Push it through a continuous f and you get a new road w(t) = f(z(t)) in the w-plane. This is how you actually draw what a map does: feed it a grid of horizontal and vertical lines and watch the image grid bend.
- Pick a familiar test shape in the z-plane — a horizontal line y = c, a vertical line x = c, or the circle |z| = r.
- Parametrize it: e.g. the circle is z(t) = r e^(i t) for t in [0, 2 pi).
- Apply the map to get w(t) = f(z(t)), and simplify into the form w = u + i v.
- Identify the resulting curve in the w-plane — is it a line, a circle, a parabola? — and that is the image of your test shape.