Linear Approximation — Approximating Values of a Function...

SECTION 1

Historical Context & Motivation

Long before calculators or computers existed, mathematicians needed ways to estimate complicated quantities like square roots, cube roots, and trigonometric values. Imagine you know that √4 = 2, but someone asks you for √4.1. Without a calculator, that's a tough question. Linear approximation arose as an elegant answer to this challenge: if you zoom in close enough to any smooth curve, it starts to look like a straight line, and straight lines are easy to work with.

The idea of replacing a curve with a nearby line has roots in the earliest days of calculus. Isaac Newton and Gottfried Wilhelm Leibniz both recognized that derivatives measure how a function changes, and that this rate of change could be used to predict function values at nearby points. Over the centuries, mathematicians refined the concept into what we now call linearization — the process of building a tangent-line approximation around a known point.

~250 BCE

Archimedes' Estimations

Archimedes used geometric methods to approximate values like π, employing the idea that curves can be approximated by simpler shapes — a precursor to linear approximation.

1665–1687

Newton Develops Calculus

Isaac Newton introduced the concept of fluxions (derivatives) and showed that a function's instantaneous rate of change could predict nearby values, forming the mathematical backbone of linear approximation.

1684

Leibniz Publishes Differential Calculus

Gottfried Wilhelm Leibniz independently developed calculus with the notation dy/dx, making it easy to express how small changes in x produce small changes in y — the essence of linearization.

1715

Taylor's Series Expansion

Brook Taylor published his theorem showing that functions can be approximated by polynomials. Linear approximation is the first-order (simplest) case of a Taylor polynomial.

The central question that linear approximation addresses is simple yet powerful: if you know the exact value of a function at one point and you know how fast it's changing there, can you estimate the function's value at a nearby point? The answer is yes — and the tool you use is the tangent line.

SECTION 2

Core Principles & Definitions

Linear approximation rests on one beautiful geometric fact: when you zoom in on a smooth, differentiable curve at any point, the curve looks more and more like a straight line. This property is called local linearity. Because straight lines are simple to evaluate, we can use the tangent line as a stand-in for the function near the point of tangency.

Local Linearity

A differentiable function, when viewed up close at a point, behaves almost exactly like a straight line. The closer you stay to the point, the better the straight-line approximation.

Tangent Line as Approximator

The tangent line at a point shares the same value and the same slope as the function at that point. This makes it the best possible linear approximation near that location.

Linearization L(x)

The linearization of f at x = a is the function L(x) = f(a) + f ′(a)(x − a). It's simply the equation of the tangent line, repackaged as an approximation tool.

Accuracy Depends on Distance

The approximation is most accurate when x is close to a. As you move farther from the base point, the curve and the tangent line diverge, and the error grows.

✦ KEY TAKEAWAY

Think of linear approximation like using a straight ruler to trace a gentle hill. If you place the ruler so it just touches the hilltop, the ruler's edge closely follows the hill's surface nearby. But if you extend the ruler out far enough, the hill curves away and the ruler no longer matches. The tangent line is your ruler, the hill is your function, and staying close to the point of contact is the key to an accurate estimate.

SECTION 3

Visual Explanation — Zooming In on a Curve

The diagram below shows a curve f(x) and its tangent line at the point x = a. Notice how the tangent line hugs the curve near the point of tangency but drifts away as you move farther from a. The vertical gap between the curve and the tangent line represents the approximation error — small near a, larger as you move away.

The violet curve is the actual function f(x). The dashed cyan line is the tangent line L(x) at the point (a, f(a)). The red segment shows the error between the true value f(x) and the approximation L(x). Notice how this error grows as x moves away from a.

In the diagram above, the tangent line and the curve share the same point and the same slope at x = a. That's why the linearization L(x) gives a great estimate when x is near a. But as x drifts to the right (or left), the curve bends away from the line, and the red error segment grows. This visual captures the central trade-off: linear approximation is powerful but local.

SECTION 4

Mathematical Framework

The mathematical foundation of linear approximation comes directly from the equation of a tangent line. Recall from earlier math classes that the equation of a line through a point (a, f(a)) with slope m is y − f(a) = m(x − a). In calculus, the slope of the tangent line at x = a is the derivative f ′(a). Solving for y gives us the linearization formula.

LINEARIZATION FORMULA

L(x) = f(a) + f ′(a) · (x − a)

Where a is the base point (a value where f is easy to compute), f(a) is the known function value, f ′(a) is the derivative at the base point, and x is the nearby value you want to estimate.

You can also express this in terms of a small change. Let Δx = x − a represent how far x is from the base point. Then the approximation becomes:

DIFFERENTIAL FORM

f(a + Δx) ≈ f(a) + f ′(a) · Δx

This says: the new value ≈ the old value + (rate of change) × (small step). This form is especially useful when you think of Δx as a small nudge away from a known point.

There is a closely related concept called the differential. If we define dx = Δx (a small change in x) and dy = f ′(x) · dx, then dy is the change along the tangent line. The actual change in the function is Δy = f(x + dx) − f(x). Linear approximation says that dy ≈ Δy when dx is small.

DIFFERENTIAL

dy = f ′(x) · dx

Here dy represents the approximate change in y predicted by the tangent line, while dx is the small change in x.

💡 Choosing the Base Point

When using linear approximation, always pick a base point a where you can easily compute both f(a) and f ′(a). For example, to approximate √4.1, choose a = 4 because √4 = 2 is simple. To approximate sin(0.05), choose a = 0 because sin(0) = 0 and cos(0) = 1 are well known.

SECTION 5

Understanding the Error — When Is the Approximation Good?

Linear approximation doesn't give exact answers — it gives estimates. Understanding when those estimates are good (and when they're not) is just as important as knowing the formula. Two main factors control the accuracy: how far x is from the base point a, and how much the function curves near a.

Left panel: a function with low curvature stays close to its tangent line even relatively far from the base point a. Right panel: a function with high curvature bends sharply away, producing a much larger error at the same distance from a.

The second derivative f ″(a) measures curvature. When |f ″(a)| is small, the function bends gently and the tangent line stays close — your approximation is quite accurate. When |f ″(a)| is large, the function curves sharply, and even a small step away from a can produce a noticeable error.

Factors affecting the accuracy of a linear approximation

Factor	Good Approximation	Poor Approximation
Distance \|x − a\|	x is very close to a (small Δx)	x is far from a (large Δx)
Curvature \|f ″(a)\|	Small — the function bends gently	Large — the function bends sharply
Concavity direction	If f ″ > 0, L(x) underestimates (curve above line)	If f ″ < 0, L(x) overestimates (curve below line)

⚠️ Over or Under?

When the function is concave up (f ″ > 0), the tangent line lies below the curve, so L(x) will be an underestimate. When the function is concave down (f ″ < 0), the tangent line lies above the curve, so L(x) will be an overestimate. Knowing this helps you judge the direction of your error.

SECTION 6

Worked Example — Approximating √4.1

Let's walk through a classic example step by step. We want to use linear approximation to estimate √4.1 without a calculator.

Approximating √4.1 Using Linearization

Step 1 — Define the Function and Choose the Base PointWe want to approximate √4.1, so our function is f(x) = √x = x^1/2. We need a base point a where f(a) is easy to compute. Since 4.1 is close to 4, and √4 = 2, we choose a = 4.

f(x) = √x, a = 4, f(4) = 2

Step 2 — Find the DerivativeDifferentiate f(x) = x^1/2 using the power rule: f ′(x) = (1/2)x^−1/2 = 1/(2√x).

f ′(x) = 1/(2√x)

Step 3 — Evaluate the Derivative at the Base PointPlug a = 4 into the derivative: f ′(4) = 1/(2√4) = 1/(2 × 2) = 1/4.

f ′(4) = 1/4 = 0.25

Step 4 — Write the LinearizationUsing L(x) = f(a) + f ′(a)(x − a), we get L(x) = 2 + (1/4)(x − 4). This is the equation of the tangent line to √x at x = 4.

L(x) = 2 + (1/4)(x − 4)

Step 5 — Plug In x = 4.1Substitute x = 4.1: L(4.1) = 2 + (1/4)(4.1 − 4) = 2 + (1/4)(0.1) = 2 + 0.025 = 2.025.

√4.1 ≈ 2.025

Step 6 — Check AccuracyA calculator gives √4.1 ≈ 2.02485. Our estimate of 2.025 is off by only about 0.00015, which is an error of roughly 0.007%. That's remarkably close, achieved with simple arithmetic!

Actual: 2.02485… | Error: ≈ 0.007%

✦ WHY IT WORKS SO WELL

Notice the pattern in Step 5: the answer comes from adding a small correction (0.025) to a known value (2). The derivative tells you the rate of change, and multiplying by the small step Δx = 0.1 gives the estimated change. This "known value + small correction" pattern is the heart of linear approximation.

SECTION 7

Strengths and Limitations

Linear approximation is a powerful tool, but like any tool it has its ideal uses and its limitations. Understanding both helps you know when to reach for it and when a more precise method might be needed.

Comparing the strengths and limitations of linear approximation

Strengths	Limitations
Quick mental or pencil-and-paper estimates for complicated functions (roots, trig, logs)	Only accurate near the base point a — not useful for values far from a
Only requires knowing f(a) and f ′(a) — two values you can often compute by hand	Does not account for curvature — ignores f ″(a) and higher-order bending
Provides insight into how sensitive a function is to small changes in input	Fails at points where f is not differentiable (corners, cusps, vertical tangents)
Foundational idea for error analysis, physics formulas, and engineering approximations	Doesn't tell you exactly how large the error is without additional analysis

✦ WHEN TO USE LINEAR APPROXIMATION

Linear approximation is ideal for quick estimates when you need a "good enough" answer, or when you want to understand how a small change in input affects the output. It's the go-to tool in physics and engineering for simplifying formulas — for instance, physicists use sin(θ) ≈ θ for small angles. If you need higher precision, the natural next step is a Taylor polynomial, which adds curvature terms to improve accuracy.

SECTION 8

Connection to Taylor Polynomials and Advanced Approximation

Linear approximation is the simplest member of a larger family of approximation techniques. If using a straight line gives a decent estimate near a point, imagine how much better you'd do by using a parabola that matches both the slope and the curvature of the original function. That's exactly what a second-degree Taylor polynomial does.

Linear vs. quadratic approximation comparison

Feature	Linear Approximation (1st degree)	Quadratic Approximation (2nd degree)
Formula	f(a) + f ′(a)(x − a)	f(a) + f ′(a)(x − a) + [f ″(a)/2](x − a)²
Matches at x = a	Value and first derivative	Value, first derivative, and second derivative
Shape	Straight line (no bending)	Parabola (captures curvature)
Accuracy	Good for very small Δx	Good for a wider range around a
Computation	Needs f(a) and f ′(a)	Also needs f ″(a)

The general pattern continues: a third-degree Taylor polynomial matches the function's value, slope, curvature, and rate of change of curvature, and so on. As you add more terms, the polynomial approximation hugs the original function over a wider range. Linear approximation is the starting point of this beautiful chain of ideas, and understanding it thoroughly prepares you for Taylor series — one of the most powerful tools in all of mathematics.

SECTION 9

Practice Problems

PROBLEM 1 — CONCEPTUAL

Explain in your own words why the tangent line at a point gives a good approximation of a function near that point. What property of differentiable functions makes this possible?

PROBLEM 2 — BASIC CALCULATION

Use linear approximation to estimate √9.05. Choose an appropriate base point, write the linearization, and compute the estimate.

PROBLEM 3 — INTERMEDIATE

Use the linearization of f(x) = sin(x) at a = 0 to approximate sin(0.1). Then determine whether your estimate is an overestimate or an underestimate by examining the concavity of sin(x) near x = 0.

PROBLEM 4 — APPLIED

A spherical balloon has radius r = 5 cm. The volume of a sphere is V(r) = (4/3)πr³. If the radius increases by 0.02 cm, use the differential dV to estimate the change in volume.

PROBLEM 5 — CRITICAL THINKING

A student uses the linearization of f(x) = x³ at a = 1 to approximate 1.5³ and gets L(1.5) = 2.5. The actual value is 3.375. Explain why the approximation is so far off in this case. What could the student do to get a much better linear approximation of 1.5³?

SUMMARY

Linear Approximation — Key Concepts Review

Linear approximation exploits the property of local linearity: near any point on a smooth curve, the function behaves almost like a straight line. The linearization formula L(x) = f(a) + f ′(a)(x − a) gives the equation of the tangent line at the base point a, which serves as the approximating function. To use it, choose a base point a where f(a) and f ′(a) are easy to compute, then plug in the nearby value of x.

The accuracy of the approximation depends on two factors: the distance |x − a| from the base point and the curvature |f ″(a)| of the function. The differential dy = f ′(x) · dx estimates the change in y along the tangent line. When a function is concave up, the tangent line underestimates; when concave down, it overestimates. Linear approximation is the first-degree case of the more general Taylor polynomial, which adds higher-order terms for greater accuracy over a wider range.