Plain words: think of z as an address on a flat map. You walk x steps east and y steps north. That landing spot is the number. The set C is simply "all possible addresses on the map."
The picture: every complex number is a dot in a plane. The horizontal axis holds the "real part" x (written Rez), the vertical axis holds the "imaginary part" y (written Imz).
Figure s01 — the complex plane: an amber arrow runs from the origin to the dot z=x+iy; a dashed cyan line drops to the horizontal axis marking x=Rez, another runs left to the vertical axis marking y=Imz.
Why the topic needs it: Cauchy's theorem integrates over a path in this plane. If z were just a number on a line, there would be no loops to walk around and nothing to prove. The whole subject exists because C is two-dimensional.
The picture: draw the arrow from the origin 0 to the dot z. Its length is ∣z∣; its tilt is θ. That is the same right-triangle idea used for real vectors — x is the adjacent side, y the opposite side.
Figure s02 — the unit circle in cyan; an amber arrow of length 1 at angle θ lands on the point eiθ=cosθ+isinθ. A small white arc at the origin marks the angle θ.
Why the topic needs it: the integral formula parametrises a tiny circle as z=a+εeiθ. That notation is pure polar: ε is the modulus (radius) and θ sweeps the argument all the way around. Without polar form you cannot even write the key step.
Plain words: feed in one dot, get out another dot. The whole plane gets picked up, stretched, twisted, and set back down.
Why the topic needs it: the parent writes f=u+iv and dz=dx+idy, then splits the loop integral into a real part and an imaginary part. That split is this u,v decomposition combined with dz=dx+idy. You cannot follow the Green's-theorem step without both.
The picture: stand at a point z0. Take a tiny step h to a nearby point. Ask: how much did the output move, per unit of input step? That ratio hf(z0+h)−f(z0) is the local stretch-and-turn factor.
Figure s03 — the point z0 (amber dot) with eight cyan arrows h fanning out in all directions; the caption of the figure stresses that the limit must give one and the same value along every one of them.
Why the topic needs it — and why this tool: In real calculus a derivative only checks left and right (two directions). Here h lives in a plane, so there are infinitely many approach directions. Demanding one answer for all of them is a colossal constraint. That constraint is what forces the Cauchy–Riemann equations, which are what make loop integrals vanish. We use a limit because "instantaneous stretch factor" has no meaning without shrinking the step to zero.
The picture:u(x,y) is a height landscape over the plane. Walk due east — how steeply does the ground rise? That is ∂u/∂x. Walk due north — that is ∂u/∂y.
Why the topic needs it: the horizontal approach in §5 uses ∂x; the vertical approach uses ∂y. Matching the two directional limits term-by-term produces the Cauchy–Riemann equationsux=vy,uy=−vx (the first pairs the real parts, the second the imaginary parts). Partial derivatives are the language those equations are written in.
The picture: a bent, possibly looping arrow-path drawn on the complex plane. The little arrowhead shows travel direction.
Why the topic needs it:∮γ means "add up along the loop γ." The circle symbol on the integral sign, ∮, specifically means the path closes on itself. Everything Cauchy proves is about these loops.
The picture: at every point on the path, an arrow f(z) tells you a local push; dz is the direction you step. Their product is one contribution; the integral is the grand total of all contributions around the loop.
Figure s04 — a cyan closed loop γ with a direction arrowhead (counter-clockwise); six amber arrows tangent to the loop mark the tiny steps dz, one at each sample point, illustrating "sum f(z) times each dz around the loop."
Why the topic needs it: this is the object Cauchy's theorem sets to zero. Understanding it as "weighted sum of tiny steps" is what makes "the total cancels to 0" believable rather than magic.
The picture: a solid pancake (fine) versus a pancake with a bite missing (a loop hugging the bite is trapped).
Why the topic needs it: the Green's-theorem step demands f be holomorphic on the entire filled interior of γ. A single singularity inside is a "hole" that breaks simple connectivity, and the proof collapses there. This is exactly why ∮dz/z is not zero — the interior has a singularity at the origin.
Each box is a foundation. Read every arrow as the sentence "you need the box at the tail before the box at the head makes sense." Follow the arrows from the top boxes down; everything eventually feeds the bottom box — the parent topic.
Every arrow says "you need the left box before the right box makes sense." The three streams — the algebra of C, the calculus of f, and the geometry of paths — all merge at the parent topic Cauchy's integral theorem and formula.
Cauchy-Riemann equations are born from §5 (the two directional limits) + §6 (partials).
Green's theorem is what converts the loop integral of §8 into an area integral that the CR equations kill.
Once loops vanish, deforming a contour to a tiny circle gives the integral formula, and repeating the trick with higher powers of (z−a) gives the Residue theorem.
The rigidity you glimpse in §5 later powers Liouville's theorem and the Maximum modulus principle.