Before you can follow a single line of Derivatives of eˣ and aˣ — proofs, you need to be fluent in the symbols it throws around without pausing. This page builds each one from nothing, in the order they depend on each other. Read top to bottom; nothing here uses a symbol defined lower down.
The picture: a number line for the inputs x at the bottom, and for every input a height plotted above it. Join all those heights and you get a curve — the graph of the function.
Why the topic needs it: the whole chapter asks "how steep is this curve?", so we first need the curve itself. In our case the machine is f(x)=ax or f(x)=ex.
Why the topic needs it: this single splitting trick is the reason the derivative of an exponential comes back as itself. In the parent proof, ax+h=axah lets the fixed part ax slide out of the limit, leaving a pure number behind.
The picture: pick two points on the line. Draw a horizontal step (the run) and the vertical step it forces (the rise). Their ratio is the slope, and it's the same everywhere on a straight line.
Why the topic needs it: "how steep is the exponential?" is the question "what is its slope?" But a curve's steepness changes from point to point — so we need one more idea.
Why the topic needs it: every derivative in the parent note is a limit, and the special constant of §7 is defined by a limit. We introduce it now, before any formula uses it.
How we compute it (first principles): pick your point x, and a nearby point x+h a tiny run h away. The line through those two points (a secant) has slope
hf(x+h)−f(x)=runrise.
Now let the run h shrink toward 0 — using the limit symbol just defined in §5 — the second point slides into the first, and the secant swings until it becomes the tangent. That final slope is the derivative, written f′(x).
Why the topic needs it: this formula is the entry gate of the entire proof. Everything the parent note derives starts by feeding ax into this machine.
Why the topic needs it: the trick a=elna rewrites every exponential in terms of e, so the one clean rule (ex is self-derivative) can crack them all. The leftover factor turns out to be lna — the true identity of the mystery constant M(a) from §7.
Why the topic needs it: after writing ax=e(lna)x, the inner function is u=(lna)x with slope u′=lna, and the outer is g(u)=eu with g′(u)=eu. The chain rule pulls out that lna — delivering dxdax=axlna.
Read the map like a river flowing downhill: the sources at the top (a function, powers, slope, the limit) merge into the first-principles derivative, which produces the constant M(a); choosing the base that makes M(a)=1definese and gives the self-derivative on the left branch, while rewriting any base through lna and applying the chain rule gives the general rule on the right branch. Both branches land on the two boxed results of the parent note.