the topological degree of a map

Stretch a rubber band around a fixed peg and let go after deforming it; the number of times it wraps the peg, counting direction, cannot change as long as you never let it cross the peg. That wrap count is a topological degree — a whole-number 'how many times around' that is stable under continuous deformation. In complex analysis it is exactly the winding number of an image curve, and it gives the counting in the argument principle a topological soul.

For a holomorphic map f and a closed contour gamma avoiding the zeros of f, the topological degree of the map z mapsto f(z), measured at the value 0, is the winding number of the image curve f(gamma) around the origin: the net number of counterclockwise turns. By the argument principle this integer equals (zeros) minus (poles) of f inside gamma. The key fact is invariance: if you deform gamma, or deform f continuously, without ever letting the image curve pass through 0, the degree cannot jump — it is a discrete integer and continuous deformation cannot change an integer by a fractional amount, so it stays fixed. This is why the count is robust: small perturbations of f leave the number of enclosed zeros alone, until a zero crosses the boundary.

This viewpoint unifies a lot. Rouche's theorem is precisely the statement that a small enough perturbation does not change the degree (the image curves stay on the same side of 0). The fundamental theorem of algebra becomes a degree argument: a degree-n polynomial maps a large circle to a curve winding n times around 0, so it must have n roots inside. The same notion of degree extends far beyond complex analysis — to continuous maps between spheres and manifolds in topology — but here it has a clean analytic formula as a contour integral. The honest point: degree counts with sign and multiplicity, so it can hide cancellation if you forget that poles count negatively.