Intuition The ONE core idea
The whole power rule is nothing but one honest question — "when I nudge the input x by a hair, how fast does the output x n move?" — answered with careful algebra. Every symbol below exists only to make that nudge, measure the movement, and shrink the nudge to zero cleanly.
Before you can read the parent proof, you need to own every mark on the page. Below, each symbol is built from nothing: a plain-words meaning, a picture, and the reason the proof cannot live without it. Read them in order — each one leans on the one before.
x — the input we wiggle
x is just a number we are free to change . Picture a slider on a ruler; wherever you park it, that is the current x . The proof asks what happens to a formula as this slider moves a tiny bit .
n — how many copies we multiply
x n means ==multiply x by itself n times==: x 3 = x ⋅ x ⋅ x . The small raised number n is the exponent (or power ); the big number x is the base .
The parent note's "tower of blocks" is exactly this: n levels, each x times wider. Look at the figure — the height of the tower is n , and its volume is x n .
Intuition Why the proof cares about
x n specifically
Almost every curve you meet — parabolas x 2 , cubes x 3 , roots x 1/2 , reciprocals x − 1 — is a power of x . If we can differentiate x n once for every kind of n , we have unlocked nearly every derivative at once.
Common mistake Base vs. exponent — do not swap them
x n has a variable base x and a constant exponent n . That is a completely different animal from n x (constant base, variable exponent, e.g. 2 x ). The power rule is only about the first one. See Derivative of exponential functions for the other.
The parent proves the rule in stages sorted by the type of n . So you must recognise each type.
Definition The families of exponents
Positive integers { 1 , 2 , 3 , … } , written Z + — "counting numbers". Picture whole, complete towers.
Zero , n = 0 . Picture no multiplication at all — by convention x 0 = 1 .
Negative integers { − 1 , − 2 , − 3 , … } . Picture a reciprocal : x − m = x m 1 , "flip it over".
Rationals q p — a whole number over a whole number, written Q . Picture roots : x 1/2 = x , x 2/3 = 3 x 2 .
Intuition Why split into these exact stages?
Because each type needs a different tool (you cannot expand x 1/2 the way you expand x 3 ). The symbol Z + literally tells you "you're allowed the binomial theorem here"; the symbol Q warns you "switch to the implicit trick."
These are the two exponent types beginners fear, so let's anchor them visually before the proof uses them.
x − m means "one over"
x − m = x m 1 . The minus in the exponent is an instruction to flip , not to make the number negative. For x = 2 : 2 − 3 = 8 1 , still positive.
x p / q means "root and power"
x p / q = ( q x ) p : take the q -th root, then raise to p . The bottom number q is how many equal factors multiply to give x ; e.g. x 1/2 = x because x ⋅ x = x .
Notice in the figure how x 1/2 (cyan) rises fast near 0 then flattens , while x 2 (amber) starts flat then steepens. That difference in steepness is exactly what a derivative measures — hold that thought.
Definition Slope — steepness of a line
Between two points, slope = run rise = change in input change in output . A big slope means the graph shoots up steeply; slope 0 means flat; a negative slope means it falls.
Intuition Why "slope" is the whole game
"How fast does the output move when I nudge the input?" is the slope question. The derivative is just the slope of a curve at a single point — the steepness of the curve's own direction right there.
But a curve's steepness changes from point to point. To pin it down at one spot we need a limit.
h — the tiny nudge
h is ==a small change we add to x ==. We look at the input x + h : the slider moved a hair to the right. h is meant to become ridiculously small , but never exactly 0 (yet).
In the figure, the amber chord connects the point at x to the point at x + h . Its steepness is the difference quotient. As h shrinks (dashed chords), the chord swings until it lies flat along the curve — that final direction is the tangent.
h → 0 lim — "what value does this head toward?"
lim h → 0 ( stuff ) asks: as h gets closer and closer to 0 (but is never plugged in as 0 ), what single number does stuff approach? We can't set h = 0 directly, because the difference quotient would read 0 0 — undefined. The limit dodges that by approaching instead of arriving .
Intuition Why a limit and not just algebra?
Every fixed h gives the slope of a chord (a shortcut across the curve), which is slightly wrong. Only the limiting chord — swung down to touch at a single point — is the true instantaneous slope. The whole reason the proof divides by h and then sends h → 0 is to reach that true tangent. Full detail lives in Limit definition of the derivative .
( x + h ) n expanded
Multiplying out ( x + h ) n for a positive integer n gives a finite sum:
( x + h ) n = x n + n x n − 1 h + ( 2 n ) x n − 2 h 2 + ⋯ + h n .
Each term trades one x for one h as you move right.
( k n ) — "choose", the counting coefficient
( k n ) (read "n choose k ") counts how many ways to pick k items from n ; here it counts how many terms collapse into each power of h . We only ever need ( 1 n ) = n — the coefficient of the first h -term — because every term after it dies in the limit.
Intuition Why THIS tool for positive integers?
The binomial theorem is the only tool that separates the "h 0 part" (x n , which cancels), the "h 1 part" (n x n − 1 h , which survives), and the "h 2 and higher" parts (which vanish). It hands the answer n x n − 1 on a plate — but only when the expansion stops , i.e. only for non-negative integer n . For n = 2 1 or n = − 3 the expansion runs forever, so we must use other tricks. See Binomial theorem .
Definition Combining fractions
A 1 − B 1 = A B B − A . We stack two fractions over one shared denominator so their tops can be compared and simplified.
Intuition Why the proof needs this
For x − m = x m 1 the difference quotient is a difference of fractions . Putting them over a common denominator ( x + h ) m x m reveals the familiar chunk x m − ( x + h ) m on top — which Stage 1 already knows how to handle. That's the whole Stage-3 idea. The Quotient rule is an alternative route to the same result.
Definition Implicit differentiation
If y depends on x but you don't have "y = something in x " cleanly, you differentiate both sides of an equation term by term, treating y as a hidden function of x . In Stage 4 we write y = x p / q , raise to power q to kill the fraction (y q = x p , now both sides are integer powers !), then differentiate.
Definition Chain rule (the part you must not forget)
When you differentiate something built from y , and y itself changes with x , you multiply by d x d y :
d x d ( y q ) = q y q − 1 ⋅ d x d y .
The extra factor d x d y is the "and y is moving too" correction.
Common mistake The classic Stage-4 slip
Writing d x d y q = q y q − 1 without the d x d y . It looks like the power rule, but y is a function of x , so the chain rule demands the extra factor. Details: Chain rule and Implicit differentiation .
Number types: positive int, zero, negative, rational
Binomial theorem and choose
Implicit diff plus chain rule
Test yourself — reveal only after answering.
What does x n mean in plain words, and which part is the base? Multiply x by itself n times; x is the base, n is the exponent.
What is x 0 by convention? 1 .
Rewrite x − 3 as a fraction. x 3 1 — the minus means "flip", not "negative".
What does x 2/3 mean as a root-and-power? The cube root of
x , squared:
( 3 x ) 2 .
Which number-set symbol says "binomial theorem is allowed"? Z + (non-negative integers) — finite expansion.
Write the difference quotient for f . h f ( x + h ) − f ( x ) .
Geometrically, what is the difference quotient? The slope of the chord joining the points at x and x + h .
Why can't we just set h = 0 in the difference quotient? It becomes 0 0 , undefined; the limit approaches 0 instead.
Definition of the derivative as a limit? f ′ ( x ) = lim h → 0 h f ( x + h ) − f ( x ) .
In ( x + h ) n , which term has coefficient ( 1 n ) = n ? The n x n − 1 h term — the one that survives the limit.
Why does the binomial proof fail for n = 2 1 ? The expansion becomes an infinite series, so you can't divide-and-cancel cleanly.
How do you combine A 1 − B 1 ? A B B − A over a common denominator.
Differentiate y q with respect to x (with y a function of x ). q y q − 1 d x d y (chain rule attaches the d x d y ).