Where −1 comes from
We have a zero with a reciprocal and a grade. The natural next thing to do with any number is raise it to powers and see what comes out. Almost every rung of that ladder is forced on us by rules we already accept — and then exactly one rung is left open. What we choose to write there is where the number −1 comes from, and with it, the whole complex plane.
Take the ordinary rules of exponents — the ones you already believe — and apply them to 0 and ω without flinching, without adding a single new assumption. Anything to the zero power is 1.2 Anything to the first power is itself. A negative power is a reciprocal, and we already know zero's reciprocal is ω. That alone pins down the entire integer ladder:
None of that is a choice. Every entry follows from rules you would insist on for any ordinary number, applied evenly to the two new poles. The grades line up too: 0ⁿ sits at grade +n, ωⁿ at grade −n, exactly as the bookkeeping demands.
Run down that ladder and you find something surprising: after the integer powers are all fixed, essentially one value is left unpinned. The rules say nothing about raising zero to the power ω itself — 0^ω — nor its mirror image. It is the one genuinely free choice in the whole construction.
But it is a free choice hemmed in on every side. Whatever we put there has to keep the powers closing consistently — reciprocals still reciprocals, the ladder still reversible, the whole system still total. Run through the constraints and one value threads all of them at once. We set it, as the natural normalization that makes everything close:1
That single line pays for the whole complex plane, and it does it without positing anything. Watch what the half-power does. If 0^ω = −1, then 0^(ω/2) is the thing that, squared, gives −1 — which is precisely the job description of the imaginary unit:
And because the powers of ω wrap rather than run off, going twice around brings you home:
So from one forced ladder and one natural normalization, three things arrive together that ordinarily cost real machinery: the number −1, the imaginary unit i, and a circle that closes exactly. The standard route to these posits i by decree, builds exp from an infinite series, and defines π through a limit. Here none of that is needed — no series, no limit, no decree beyond the one normalization we were honest about choosing.
We now have the imaginary unit sitting inside the powers of zero, and a hint that those powers trace a circle. Two questions come straight out of that. First — in what sense is i "a zero," really? And second — if the circle is exact, can we actually compute points on it, by hand, with no limits and no approximation? Those are the next two chapters.