AP Calculus AB · Lesson Review · October 7, 2026

Linear Approximation and Differentials

Use tangent lines to estimate nearby function values, choose convenient base points, compare estimates with calculator values, and connect linearization to the differentials dx and dy.

35-minute edited lesson11 chaptersCaptions + transcript

01 · Watch

Lesson video

02 · Understand

Learning targets

03 · Connect

Key ideas

Linear approximation, linearization, and tangent-line approximation describe the same local model. A differential records the corresponding change along that tangent line.

01Start with point-slope form
y−f(a)=f′(a)(x−a)y-f(a)=f'(a)(x-a)The point is (a,f(a)); the slope is f′(a). Evaluate the derivative at the fixed base point a.
02Separate the exact line from the estimate
L(x)=f(a)+f′(a)(x−a),f(x)≈L(x)L(x)=f(a)+f'(a)(x-a),\qquad f(x)\approx L(x)The first equality defines the line exactly. The approximation symbol describes how that line estimates the original function near a. They agree exactly at a.
03Choose an easy, nearby base point
9.3:a=9,ln⁡(1.2):a=1\sqrt{9.3}:a=9,\qquad\ln(1.2):a=1Select a point where the function value and derivative are easy to evaluate. Staying close to a matters more than making the line easy to use far away.
04Radian measure is essential for the sine approximation
sin⁡x≈xnear x=0\sin x\approx x\qquad\text{near }x=0At zero, sine has value zero and slope one when x is measured in radians. The degree-measure version has a different slope.
05Differential versus actual change
dy=f′(a) dx,Δy=f(a+dx)−f(a)dy=f'(a)\,dx,\qquad\Delta y=f(a+dx)-f(a)dx is the chosen input change. dy is the change predicted by the tangent line; Δy is the actual change on the curve. For a small dx, Δy≈dy.
06Concavity can explain the direction of error
f′′>0⇒L≤f,f′′<0⇒L≥ff''>0\Rightarrow L\le f,\qquad f''<0\Rightarrow L\ge fWhen these concavity conditions hold throughout the relevant interval, a tangent line lies below a concave-up curve and above a concave-down curve. This is an additional way to check the examples.

04 · Apply

Worked examples

These examples follow the methods developed in the recorded lesson.

Original polynomial

Replace a complicated polynomial near zero

f(x)=x6−7x5+4x4−x3+3x2+2x−8f(x)=x^6-7x^5+4x^4-x^3+3x^2+2x-8

The recording uses this polynomial to show why a simpler local model can be useful.

f(0)=−8,f′(0)=2f(0)=-8,\qquad f'(0)=2

Differentiate first, then evaluate at the base point.

L0(x)=−8+2x,f(0.1)≈L0(0.1)=−7.8L_0(x)=-8+2x,\qquad f(0.1)\approx L_0(0.1)=-7.8

This estimate avoids substituting 0.1 into all seven terms. It is an approximation, not an exact polynomial value.

A different base point

The same function gives a different local line at one

f(1)=−6,f′(1)=6−35+16−3+6+2=−8f(1)=-6,\qquad f'(1)=6-35+16-3+6+2=-8

Use the original polynomial from the preceding example.

L1(x)=−6−8(x−1)=−8x+2L_1(x)=-6-8(x-1)=-8x+2

This line is anchored at x=1. It differs from the tangent line at zero.

f(0.9)≈L1(0.9)=−8(0.9)+2=−5.2f(0.9)\approx L_1(0.9)=-8(0.9)+2=-5.2

Use the arithmetic correction made during the recording. The actual polynomial value is −5.476589. This line is useful near one; the line at zero models a different neighborhood.

Sine near zero

Use the familiar small-angle model

f(x)=sin⁡x,f(0)=0,f′(0)=cos⁡0=1f(x)=\sin x,\qquad f(0)=0,\qquad f'(0)=\cos0=1

Use radians.

L(x)=x,sin⁡(0.2)≈0.2L(x)=x,\qquad\sin(0.2)\approx0.2

The calculator comparison in the lesson gives about 0.198669 for the actual value. The approximation is slightly high.

Logarithm near one

Estimate a logarithm without a calculator

f(x)=ln⁡x,f(1)=0,f′(1)=1f(x)=\ln x,\qquad f(1)=0,\qquad f'(1)=1

The real logarithm requires x>0. One is a convenient nearby base point.

L(x)=x−1,ln⁡(1.2)≈0.2L(x)=x-1,\qquad\ln(1.2)\approx0.2

The actual value is about 0.182322. Since ln x is concave down, its tangent line gives an overestimate.

Square root near nine

Use a nearby perfect square

f(x)=x,f(9)=3,f′(9)=129=16f(x)=\sqrt{x},\qquad f(9)=3,\qquad f'(9)=\frac1{2\sqrt9}=\frac16

The derivative formula requires x>0.

L(x)=3+16(x−9)L(x)=3+\frac16(x-9)

The base point is nine, not 9.3.

9.3≈3+0.36=3.05\sqrt{9.3}\approx3+\frac{0.3}{6}=3.05

The actual value is about 3.049590, so the estimate is slightly high.

Original differential example

Find the differential of sine

y=sin⁡x,dydx=cos⁡xy=\sin x,\qquad\frac{dy}{dx}=\cos x

The derivative describes the slope.

dy=cos⁡x dxdy=\cos x\,dx

The differential describes the tangent-line change for a chosen input increment dx. It is generally an estimate for the corresponding actual change in sine.

dx=dycos⁡x(cos⁡x≠0)dx=\frac{dy}{\cos x}\qquad(\cos x\ne0)

Solving this differential relation for dx requires a nonzero slope. At a point where cosine is zero, this division cannot be used.

05 · Avoid

Common mistakes

Using a variable slope in a fixed line.

Evaluate f′ at a. The expression f′(x)(x−a)+f(a) is generally not the linearization at a.

Confusing a with the input to estimate.

For √9.3, choose a=9, then substitute x=9.3 into the resulting line.

Treating an approximation as an equality.

Write L(x)=f(a)+f′(a)(x−a), but use f(x)≈L(x) for a nearby estimate. A line can be a poor model far from its base point.

Mixing degrees, radians, and changes.

The sine approximation x uses radians. Keep dx, dy, and the actual change Δy distinct; dy is not generally the exact change on the curve.

06 · Check yourself

Try it first

Use these five fresh problems to check your understanding. Open a solution after you try it.

1Linearize at a=2, then estimate the nearby value: f(x)=x3,f(2.03)f(x)=x^3,\qquad f(2.03)

f(2)=8,f′(2)=12,L(x)=8+12(x−2)f(2)=8,\qquad f'(2)=12,\qquad L(x)=8+12(x-2)

Evaluate both the function and its derivative at two.

(2.03)3≈8+12(0.03)=8.36(2.03)^3\approx8+12(0.03)=8.36

Because x³ is concave up near two, this tangent estimate is slightly low.

2Choose a convenient base point and estimate: 16.4\sqrt{16.4}

a=16,f(16)=4,f′(16)=18a=16,\qquad f(16)=4,\qquad f'(16)=\frac18

Use the nearby perfect square.

16.4≈4+0.48=4.05\sqrt{16.4}\approx4+\frac{0.4}{8}=4.05

The square-root function is concave down here, so this is an overestimate.

3Use the tangent line at a=1 to estimate: ln⁡(0.96)\ln(0.96)

L(x)=x−1,ln⁡(0.96)≈−0.04L(x)=x-1,\qquad\ln(0.96)\approx-0.04

The input change is −0.04. The estimate should be negative because 0.96 is less than one.

4Compare the differential with the actual change: y=x2,x=3,dx=0.1y=x^2,\qquad x=3,\qquad dx=0.1

dy=2(3)(0.1)=0.6dy=2(3)(0.1)=0.6

Use the tangent-line slope at the starting input.

Δy=(3.1)2−32=0.61\Delta y=(3.1)^2-3^2=0.61

The differential approximates the actual change; their difference is 0.01.

5Use the given local data to estimate the function value: f(2)=5,f′(2)=−3,f(1.98)f(2)=5,\qquad f'(2)=-3,\qquad f(1.98)

L(x)=5−3(x−2)L(x)=5-3(x-2)

The original function need not be supplied to build its local line.

f(1.98)≈5−3(−0.02)=5.06f(1.98)\approx5-3(-0.02)=5.06

A negative input change times a negative slope produces a positive predicted output change.

07 · Revisit

Transcript and captions

Search the corrected transcript for a method or worked example, or download the caption files above.

Download corrected transcript