A zero can be gentle or forceful
Back on the power-series rung you met the zeros of an analytic function and learned that they are isolated — each one sits alone, with no other zeros crowding in. But not all zeros are alike. Picture the graph of a function pressing down toward zero: it might cross cleanly, like z passing through the origin, or it might flatten against zero, touching and curving away, like z^2. The order of a zero is the one whole number that captures exactly how hard the function presses down. A simple zero (order 1) is a clean crossing; a higher-order zero is a tangency where the function lies flatter and flatter against zero.
There are three ways to say the same thing, and it pays to hold all three at once. First, by factoring: z_0 is a zero of order m if you can write f(z) = (z - z_0)^m g(z) with g analytic and g(z_0) not equal to 0 — exactly m factors of (z - z_0) hide inside f. Second, by the Taylor series: the expansion at z_0 begins at the m-th term, f(z) = a_m (z - z_0)^m + a_(m+1)(z - z_0)^(m+1) + ..., with a_m not equal to 0. Third, by derivatives: f(z_0) = 0, f'(z_0) = 0, ..., up through f^(m-1)(z_0) = 0, but f^(m)(z_0) is not 0. The function and its first m - 1 derivatives all vanish together, and the m-th is the first to survive.
The practical recipe falls right out of the second description: expand the Taylor series and read off the index of the first surviving coefficient — that index is m. Take 1 - cos z at the origin. We know cos z = 1 - z^2/2 + z^4/24 - ..., so 1 - cos z = z^2/2 - z^4/24 + ..., and the series first stirs to life at z^2. The order is 2. No derivatives to grind through; the series simply told us.
A pole can be a low spike or a towering one
Now mirror the whole story upside down. In the last guide you classified an isolated singularity and met the pole as the case where the function blows up to infinity in a controlled way — the principal part of the Laurent series has only finitely many negative terms. But not all poles blow up equally fast. 1/(z - z_0) sails off to infinity; 1/(z - z_0)^2 races there far faster; 1/(z - z_0)^5 faster still. The order of a pole is the whole number that measures how violently the spike climbs — how steep it is.
The cleanest definition reads straight off the Laurent series: the order is the depth of the deepest negative term. If f(z) = a_(-m)/(z - z_0)^m + ... + a_(-1)/(z - z_0) + (analytic part) with a_(-m) not equal to 0, the pole has order m. A pole of order 1 is called simple; orders 2, 3, ... are double, triple, and so on. Notice the contrast with a zero, where order counts how many factors of (z - z_0) you can pull out; for a pole, order counts how many factors of (z - z_0) you would have to multiply back in to clear the denominator.
Zeros and poles are two faces of one coin
Here is the idea that makes orders genuinely powerful, and it is the heart of this guide. Reciprocating turns a zero into a pole and a pole into a zero — and it does so order for order. This is the content of the result on the zeros and poles of 1/f: if f has a zero of order m at z_0, then 1/f has a pole of order m there, and conversely. The same whole number simply flips sides. A clean crossing of zero (order 1) becomes a simple pole; a flat double zero becomes a double pole.
Why is this true? Use the factored picture. If f(z) = (z - z_0)^m g(z) with g analytic and g(z_0) not equal to 0, then 1/f(z) = 1/(z - z_0)^m times 1/g(z). Since g(z_0) is nonzero, 1/g is analytic and nonzero right at z_0, so it contributes no singularity of its own — it just rides along as a well-behaved factor. All the drama lives in that 1/(z - z_0)^m out front, which is precisely a pole of order m. The factorisation that named the zero's order hands you the pole's order for free.
This duality is also how you read off a pole when a function arrives as a quotient. Suppose f = g/h with g and h holomorphic and g(z_0) not equal to 0. Then the order of the pole of f at z_0 equals the order of the zero of h at z_0 — the numerator is just along for the ride. So to find the pole, you forget the top and ask only how hard the bottom vanishes.
Orders do arithmetic
The real joy of orders is that they obey simple bookkeeping, which lets you read off the behaviour of a complicated expression by inspection. When you multiply functions, their orders add at a shared point: if f has a zero of order j at z_0 and g has a zero of order k there, then fg has a zero of order j + k. Squaring a function doubles the order of every zero. And a zero in the numerator cancels pole order in the denominator: a zero of order k upstairs erases k units of pole order downstairs, possibly leaving a lower-order pole, a removable point, or even a zero, depending on which side wins.
A worked example pins this down. Take f(z) = 1/(z^2 sin z) at z_0 = 0. The factor z^2 contributes a zero of order 2 in the denominator. And sin z = z - z^3/6 + ... has a simple zero at 0 — order 1. So the denominator z^2 sin z vanishes to total order 2 + 1 = 3, and by the quotient rule for poles, f has a pole of order 3. Check it with the shortcut: z^3 f(z) = z^3/(z^2 sin z) = z/sin z, which tends to 1 as z goes to 0 — a finite nonzero limit, confirming the order is exactly 3, no more and no less.
f(z) = g(z) / h(z), g(z_0) != 0
zero of h at z_0 has order m -> f has a POLE of order m
zero of g at z_0 has order k -> contributes a ZERO of order k
net order at z_0 = (order of zero of g) - (order of zero of h)
positive -> a zero of that order
negative -> a pole of that order
zero -> removable, with a finite nonzero valueWhy the order is worth knowing
Order is not a curiosity to memorise — it is the number the rest of the subject keeps reaching for. The most immediate payoff is computing a residue, the single Laurent coefficient that unlocks the residue theorem. The formula you use depends entirely on the order: a simple pole gives the residue by a one-line limit, while a pole of order m needs m - 1 derivatives. Know the order first, and you know which tool to pick up.
Order is also what makes counting honest. A degree-n polynomial has exactly n zeros only when each is counted according to its order, and the argument principle counts zeros and poles by their orders inside a contour. Geometrically, near a zero of order m the function behaves like (z - z_0)^m, so it wraps angles by a factor of m and is locally m-to-1 — the order is literally how many times the map folds the neighbourhood onto itself. The bookkeeping of orders, flowing freely between zeros and poles, is exactly what makes the meromorphic functions of the next guide so workable: such a function is described, up to detail, by the locations and orders of its zeros and poles.