Complex analysis and contour reasoning
Use analyticity, singularities, and contour topology to evaluate integrals with defensible assumptions.
On this page
The question for today
What you’ll understand
- Explain why complex differentiability is stronger than real differentiability.
- Evaluate a contour integral from the singularities it actually encloses.
- Track branch choices and contour prescriptions rather than hiding them in notation.
Complex analysis turns local differentiability into remarkable global information. Knowing a function around a closed curve can determine its values inside, and an integral can be controlled by a few singularities. This machinery appears in wave propagation, stability theory, and quantum propagators. Its power depends on analyticity in a specified region and on the topology of the contour. A diagram with poles is useful only when accompanied by a careful statement of what lies inside and what has been excluded.
Differentiation must agree along every complex direction
For a complex function f(z), the derivative must have the same limit as an increment approaches zero from any direction in the plane. Writing f=u+iv yields the Cauchy–Riemann relations between the two real component functions. When u and v have continuous first partial derivatives on an open region, these relations imply holomorphicity there. Satisfying them at one isolated point does not by itself establish analyticity on a neighborhood; the domain and regularity hypotheses cannot be discarded.
Holomorphic functions are locally represented by convergent power series. This is much stronger than real smoothness: a real function can have derivatives of all orders without equaling its Taylor series near a point. The complex conjugate f(z)=conjugate(z) makes the direction issue vivid. Along real increments its difference quotient is one, while along purely imaginary increments it is minus one. A differentiable map of two real coordinates need not be complex differentiable.
Contours detect singularities and winding
Cauchy's integral theorem says a holomorphic function has zero integral around an appropriately contractible closed contour in its domain. A hole matters. The function 1/z is holomorphic away from zero but has integral 2πi around a counterclockwise circle enclosing zero. Cauchy's integral formula sharpens this observation: integrating f(z)/(z−a) around a suitable enclosing contour recovers f(a). The local value is tied to data around the boundary.
For isolated singularities, the residue is the coefficient of (z−a)^(−1) in the Laurent expansion. The residue theorem adds these coefficients, weighted by winding number, to evaluate a closed contour integral. A simple pole can often be handled by multiplying by z−a and taking the limit. The sign of orientation, excluded poles, and contributions from added arcs all matter. A real integral obtained by closing a contour is incomplete until the extra arc is shown to vanish or its contribution is computed.
Branches and boundary values carry information
A complex logarithm cannot be single-valued and continuous on all nonzero complex numbers. Going once around zero changes the accumulated argument by 2π. A branch chooses a domain and a consistent argument there, often by removing a ray. Fractional powers inherit this choice through the logarithm. Moving across a branch cut changes a selected representation, and analytic continuation around a branch point may return a different branch. This is global information not visible in one local derivative.
In physics, poles near the integration contour require a prescription for how the contour passes them. Replacing a denominator by one shifted by a small positive or negative imaginary term encodes a limiting boundary value, not an inconsequential numerical perturbation. Similar care applies to Wick rotation: singularities, falloff, and contour deformation must permit it. Complex methods can simplify a well-defined integral, but they cannot silently choose the physical boundary conditions that made it well-defined.
Work it through
A real integral from one enclosed pole
Evaluate for real a>0.
Extend the integrand to f(z)=1/[(z−ia)(z+ia)] and close the real segment with a counterclockwise upper semicircle of radius R>a. Only the pole at ia is enclosed.
The residue is . The closed contour integral is therefore .
On the large arc, |z²+a²|≥R²−a². Arc length is πR, so the integral's magnitude is at most πR/(R²−a²), which tends to zero. This justifies discarding the arc.
Taking R→∞ leaves the requested real integral equal to π/a. The assumption a>0 both fixes the pole location and avoids a singularity on the real axis.
The integral is . Counting a pole was only part of the argument; controlling the added arc completed it.
The living notebook
Explore the model
Interactive lab · vectors
Local directions and global winding
Change the angle and magnitude of one two-dimensional vector and inspect its x/y projections. Read these as real and imaginary parts of a complex increment to compare approach directions. The diagram does not evaluate a complex derivative, integrate a contour, detect branch cuts, or prove analyticity.
- x component
- 1.393
- y component
- 0.975
- Magnitude
- 1.7
Playback traces a fixed vector and its projections. Angle and magnitude set the geometry; the moving marker does not change them. Parameter changes pause playback and recompute the model at the current elapsed time. Reset restores both parameters and returns time to zero. Playback starts paused. Adjust either parameter, then choose Play or Step.
Look forLocal derivative agreement and global winding are distinct ingredients of complex reasoning.
Pause & reason
Check your understanding
Your turn
Try it yourself
Evaluate counterclockwise. What changes if the circle has radius 1/2?
Take a moment to make a prediction or work it out on paper. Open a hint when you need one.
Give me a hint
Determine whether the simple pole is inside before applying the residue theorem.
Compare with the solution
For radius 2, z=1 is enclosed and its residue is e, so the integral is 2πie. For radius 1/2, the pole is outside and the function is holomorphic on and inside the contour, giving zero. No pole lies on either contour; a contour through z=1 would require a different formulation.
Keep the idea
Take it with you
- Holomorphicity constrains derivatives in every complex direction on an open region.
- Residues work with a domain, an orientation, and justified contour deformations.
- Branches and pole prescriptions encode global or physical choices.
Read further
Original sources and deeper explanations to follow at your own pace.
Optional progress, saved in this browser when available. Revisit any lesson at any time.
Lesson not marked complete.