A tangent line approximation replaces a curved function with the straight line that just touches it at a chosen point. This is useful because lines are much easier to calculate with than curves, especially for quick estimates. The approximation is best near the tangent point, where the function and tangent line share the same value and slope.
As you move farther away, the curve can bend away from the line, creating error.
The size of the error depends mainly on distance from the tangent point and on the curvature of the function. The second derivative measures how quickly the slope is changing, so it controls how strongly the graph bends. Taylor's theorem shows that the error is roughly proportional to the square of the distance from the tangent point.
This is why doubling the distance can make the error about four times larger when curvature stays similar.
Understanding Calculus: Tangent Line Approximation Error
The error has a direction, not just a size. When a graph bends upward, its tangent line tends to sit underneath the nearby curve. A linear estimate then gives a value that is too small.
When a graph bends downward, the line tends to sit above the curve, so the estimate is too large. This helps students check an answer before doing detailed arithmetic. The sign of the second derivative gives this local bending information.
A positive second derivative suggests an underestimate, while a negative second derivative suggests an overestimate. This conclusion is safest close to the point of contact, since a function can change its bending farther away.
A useful practical method is to find a maximum possible amount of curvature on the interval being used. Suppose the absolute value of the second derivative never exceeds some number M between the chosen input and the target input. Then the magnitude of the linear approximation error is no more than one half times M times the square of the input distance.
This is an error bound, not merely a guess. It gives a worst case limit. In science and engineering, a bound can matter more than the estimated value because it tells whether an approximation is accurate enough for the job.
Consider estimating the square root of ten by using information at nine. The square root function is easy to evaluate at nine, and its slope there can be found from calculus. The resulting line gives an estimate near ten without needing a calculator.
Since the square root graph bends downward, the estimate should be slightly too large. Its curvature is fairly small near nine, so the error is small.
This kind of reasoning appears whenever a complicated formula is evaluated near a familiar value. Examples include changes in electrical signals, small temperature corrections, population models over short times, and engineering measurements with tiny input changes.
Students should pay attention to the word near. Linear approximation is a local tool. A line can match a curve extremely well in a small neighborhood yet fail badly at a more distant input.
It can fail especially fast near places where curvature becomes large, such as near a vertical feature or a restricted endpoint. Keep track of units too. If the input is measured in meters and the output is measured in joules, the error is measured in joules.
Finally, distinguish approximation error from rounding error. Approximation error comes from replacing the function with a line.
Rounding error comes from shortening decimal values. Both can affect a final answer, but they have different causes and need different checks.
Key Facts
- Linear approximation at x = a: L(x) = f(a) + f'(a)(x - a)
- Approximation error: Error = f(x) - L(x)
- Second-order error estimate: Error ≈ (1/2)f''(a)(x - a)^2
- Taylor error form: f(x) - L(x) = (1/2)f''(c)(x - a)^2 for some c between a and x
- If f''(x) > 0, the graph is concave up and the tangent line usually lies below the curve near the point.
- If |x - a| doubles, the tangent line error is often about 4 times larger when f'' is nearly constant.
Vocabulary
- Tangent line
- A line that touches a curve at a point and has the same slope as the curve at that point.
- Linear approximation
- An estimate of a function near a point using the tangent line at that point.
- Approximation error
- The difference between the true function value and the value predicted by the tangent line.
- Second derivative
- The derivative of the derivative, which measures how the slope of a function is changing.
- Concavity
- The bending direction of a graph, determined by whether the second derivative is positive or negative.
Common Mistakes to Avoid
- Using the tangent line far from the tangent point: this is wrong because linear approximation is only reliable near the point where it is made.
- Ignoring the second derivative: this is wrong because two functions can have the same tangent line but very different errors if their curvatures are different.
- Assuming the error grows linearly with distance: this is wrong because the leading error term usually contains (x - a)^2.
- Dropping the sign of the error without thinking: this is wrong because the sign tells whether the tangent line estimate is above or below the true curve.
Practice Questions
- 1 For f(x) = x^2 at a = 3, find the tangent line approximation L(x) and compute the exact error at x = 3.2.
- 2 Use the error estimate Error ≈ (1/2)f''(a)(x - a)^2 for f(x) = sin x at a = 0. Estimate the error at x = 0.1, and compare it to the exact error sin(0.1) - 0 using a calculator.
- 3 A function has a large positive second derivative near a tangent point. Explain whether the tangent line approximation is likely to be above or below the curve nearby, and why the error grows as you move away.