← Unit 2 Guide

The Power Rule & Basic Rules

Finally, the shortcuts. After two lessons of computing derivatives with limits, this is where it gets fast: one rule handles every power of \(x\), and three more let you differentiate any polynomial term by term — no limit definition required. Plus the payoffs: horizontal tangents and the normal line.
For two lessons, finding a derivative meant setting up a limit and grinding through algebra. The power rule ends that: to differentiate \(x^n\), just bring the exponent down in front and subtract one from it — \(\frac{d}{dx}x^n = n\,x^{n-1}\). Combined with three rules so natural they barely need stating (the derivative of a constant is zero; constants factor out; sums differentiate piece by piece), you can now differentiate any polynomial in seconds.
Two lessons ago, finding the derivative of \(x^5\) meant a page of limit algebra. Watch how long it takes with the rule you're about to learn.
The limit definition
The power rule
press to race the two methods
Same answer, \(5x^4\) — but one took six lines and the other took one glance. That shortcut is the power rule, and it's the engine for everything that follows.

One rule for every power

Here it is, the rule that replaces the limit definition for any power of \(x\): bring the exponent down in front as a coefficient, then subtract one from the exponent. That's it. Slide the exponent and watch the derivative \(f'\) (gold, dashed) track the slope of \(f\) (lime) everywhere.

exponent n = 3

What you're looking at: \(f(x)=x^n\) in lime and its derivative \(f'(x)\) in gold. Notice \(f'\) is always one degree lower than \(f\).

The power rule

The boundary: the variable must be the base, not the exponent. The power rule differentiates \(x^n\) — it does not apply to \(2^x\), where \(x\) is up in the exponent.

That handles \(x^2\), \(x^{10}\), any whole-number power. But derivatives show up as roots and reciprocals too — and the same rule covers them, with one small setup step.

It works for negative and fractional powers too

The real power of the power rule: \(n\) can be any real number. Roots and reciprocals just need to be rewritten as exponents first — then the same “down and minus one” move handles them. Tap each.

Why rewrite first?

The power rule only sees \(x^r\). It can't read a fraction bar or a root symbol — so convert to a single exponent, differentiate, then convert back if the problem wants it pretty.

the moverewrite · differentiate · rewrite
One rule for a single power. Real functions are sums of powers — so we need to know what differentiation does to a sum. Happily, the answer is “nothing surprising.”

Polynomials fall apart term by term

Three rules so natural they barely need names: the derivative of a constant is zero, constants factor straight out, and the derivative of a sum is the sum of the derivatives. Together they mean you can attack a polynomial one term at a time. Step through it.

Result

Each term is independent: the \(x^4\) term doesn't care about the \(x^2\) term. The constant \(-1\) contributes nothing — its derivative is \(0\). This is why polynomial derivatives are so fast.

You keep hearing “tangent line” — the power rule gives its slope, horizontal tangents are where that slope is zero. But what is the tangent line, and how do you write its equation? Let’s pin that down before the payoffs.

The tangent line, properly

Zoom in on a smooth curve at a point and it looks like a straight line — you saw that with the magnifier last lesson. That line is the tangent: the one straight line that best matches the curve right at that point. Of all the lines you could draw through the point, the tangent is the one that hugs the curve most closely.

tilt the line

What you're looking at: a line through the point on \(f(x)=x^2\). The shaded band is the gap between line and curve. Tilt away and the gap grows; the tangent makes it smallest.

Best-fit line
total gap nearby
status

This is what “the slope at a point” really means: the tangent is the line the curve is heading along right there. Its slope is exactly \(f'(a)\) — the derivative.

So a tangent line needs just two ingredients, and you already have both: a point on the curve, \(\big(a, f(a)\big)\), and a slope at that point, \(f'(a)\). Drop them into point-slope form and you have the equation. Step through it:

point a

What you're looking at: pick where on the curve you want the tangent, then build its equation one ingredient at a time — point, slope, line.

Building the equation

The formula: \(y = f(a) + f'(a)\,(x-a)\). Read it as “start at the point's height, then rise at the tangent's slope.” This is the AP exam's most-asked tangent question, every year.

You can now differentiate any polynomial in seconds. So what is all that speed for? Here’s the first classic payoff.

Payoff 1: where the curve goes flat

Now the rule earns its keep. A horizontal tangent is a point where the slope is zero — a peak, a valley, a flat spot. Since the derivative is the slope, you find them by solving \(f'(x)=0\). Drag the point along \(f(x)=x^3-3x\) and watch the tangent flatten.

point x

What you're looking at: the tangent turns lime and level exactly where \(f'(x)=0\) — at \(x=-1\) and \(x=1\) for this curve.

Slope at the point
status

To find them by hand: set \(f'(x)=3x^2-3=0\), so \(x^2=1\), giving \(x=\pm1\). Horizontal tangents are just the zeros of the derivative.

Setting \(f'=0\) finds flat spots. The derivative also pins down the tangent’s exact slope — and from that, the line running perpendicular to it.

Payoff 2: the normal line

The normal line at a point is the line perpendicular to the tangent there. Perpendicular slopes are negative reciprocals, so if the tangent slope is \(m\), the normal slope is \(-\tfrac1m\). One derivative gives you both lines.

point x

What you're looking at: the gold tangent and the blue normal, always at right angles, on \(f(x)=0.6x^2\).

Tangent vs. normal slope

To write the normal line's equation, use the point and the slope \(-\tfrac1m\) in point-slope form. The only place this breaks is where \(m=0\) — a horizontal tangent has a vertical normal.

The whole idea on one card

The power rule: bring the exponent down in front, then subtract one from it
Constant rule: the derivative of any constant is zero (a flat line has slope 0)
Constant-multiple and sum rules: factors come out front, and you differentiate term by term
Payoffs: horizontal tangents are the zeros of \(f'\); the normal line slope is \(-1/f'\)

The ideas everything else builds on

Down and minus one

The whole power rule in four words: exponent down, minus one. \(\frac{d}{dx}x^n=nx^{n-1}\).

Rewrite first

Roots and reciprocals become exponents: \(\sqrt{x}=x^{1/2}\), \(\tfrac1{x^n}=x^{-n}\). Then the rule applies as usual.

Term by term

Differentiation is linear: constants factor out, sums split apart. A polynomial differentiates one term at a time.

Flat means f′ = 0

A horizontal tangent is where the slope is zero. Find them by solving \(f'(x)=0\) — the zeros of the derivative.

Your turn

Five problems — the power rule on whole, negative, and fractional powers, a full polynomial, and a horizontal-tangent hunt. Try each before revealing.

Quick quiz

Eight fast checks across the whole lesson.

Question 1 of 8
Score: 0 / 0

Common mistakes & exam tips

Forgetting to subtract one

\(\frac{d}{dx}x^3\) is \(3x^2\), not \(3x^3\). Bring the exponent down and reduce it. Both steps, every time.

Mishandling the constant

\(\frac{d}{dx}(7)=0\), but \(\frac{d}{dx}(7x)=7\). A lone constant vanishes; a constant times \(x\) leaves the constant behind.

Not rewriting before differentiating

You can't power-rule a fraction bar or a root symbol directly. Convert \(\frac{1}{x^2}\) to \(x^{-2}\) and \(\sqrt{x}\) to \(x^{1/2}\) first.

Using the power rule on \(2^x\)

The power rule is for a variable base with a constant exponent. When the variable is in the exponent (\(2^x\), \(e^x\)), it does not apply — that's a different rule, coming in the next lesson.

Sign slips on negative exponents

\(\frac{d}{dx}x^{-2}=-2x^{-3}\). The new exponent is \(-3\), not \(-1\) — subtracting one from \(-2\) goes down to \(-3\).

Confusing horizontal tangent with a zero of \(f\)

Horizontal tangents come from \(f'(x)=0\), not \(f(x)=0\). One is where the slope is flat; the other is where the curve crosses the axis.

On the AP exam

You can now see limits. Next: computing them — the limit laws, why plugging in usually just works, and the algebra toolkit (factoring, conjugates) for when it doesn't, including that \(\tfrac{\sqrt{x+4}-2}{x}\) mystery from the table. Computing Limits is next — or head back to the Unit 2 Guide.