Intuition for i to the power i | Ep. 9 Lockdown live math

Raising the imaginary unit to an imaginary power—specifically i^i—collapses to a real number because complex exponentials can be reinterpreted as motion with changing dynamics. The key move is rewriting i as e^{x} for a complex x, then using Euler’s formula to translate exponentials into rotations on the complex plane. Solving e^{x}=i leads to x= (π/2)i (up to adding multiples of 2πi), so i^i becomes e^{(π/2)i·i}=e^{-π/2}≈0.2079. That “imaginary-in, real-out” outcome stops looking like a trick once exponentials are treated as a system whose derivative is proportional to its current state.

The episode builds intuition in two stages. First, Euler’s formula is given a computational meaning via the power-series definition of exp(x)=1+x+x^2/2!+…, which works cleanly for complex inputs. Then the geometric meaning is reinforced: e^{iθ} corresponds to walking an angle θ around the unit circle. Under that lens, multiplying by i is a 90° rotation, so e^{(π/2)i} is exactly the rotation that lands on i.

The deeper “why” comes from dynamics. Consider e^{it} as a time-evolving position in the complex plane. Because e^{x} differentiates to itself, differentiating e^{it} introduces a factor of i; that factor rotates the position vector by 90° to produce the velocity vector. This rule—velocity equals a quarter-turn rotation of position—forces circular motion. The time needed to reach i is therefore π/2, matching the angle interpretation.

But the expression i^i introduces a second layer: raising e^{it} to the power i changes the dynamics from circular motion to exponential decay. After the transformation, the derivative becomes negative times the current value, so the motion heads toward the origin with a shrinking step size—an exponential approach to zero. After waiting π/2 units of time under this decay rule, the position lands at e^{-π/2}. In this sense, “two 90° rotations” don’t literally add angles; one rotation sets the circular dynamics, while the other rotation (through the exponent i) flips the system into decay.

A major complication then appears: e^{x}=i has infinitely many solutions because the complex logarithm is multivalued. Adding 2πi to x doesn’t change e^{x}, but it changes the resulting exponential function. That’s why values like x= (π/2)i, (5π/2)i, or -(3π/2)i all satisfy e^{x}=i yet produce wildly different outputs for e^{x·i}. The episode connects this to the general ambiguity of roots in complex numbers and to the idea of choosing a “branch” of a multivalued function.

To unify everything, the episode argues that exponentials are best expressed as exp(r x) rather than b^x when complex numbers enter. Different choices of r correspond to different branches, and varying r changes the function even if exp(r)=b stays the same. The lecture ends by extending the theme to power towers (iterating exponentiation) where ambiguity and numerical iteration can lead to cycles and even chaotic behavior, depending on the chosen branch. The central takeaway is that complex exponentiation is not just “more complicated arithmetic”—it is a family of functions whose behavior depends on how the logarithm is chosen.