Local linearity and magnification

Textbook reading

The problem that opens the lesson

Zoom around $x=1$ on $y=x^2$ and build the tangent-line approximation from the predicted slope $2$ .

Solution

Begin by identifying the mathematical object and the information that fixes it. Identify the anchor point and slope, write the line, compare exact and approximate values at several distances, and describe the error trend. The relevant conditions are not optional bookkeeping: A graph appearing linear at one screen resolution is visual evidence, not a proof of differentiability. Following that structure gives $L(x)=1+2(x-1)=2x-1$ .

Why this works

Corners, cusps, discontinuities, and violent oscillation can defeat a single local line. The calculation and the representation should agree, so a graph, diagram, table, or substitution check should support the same conclusion.

Textbook reading

What this lesson is really about

Local linearity is the tendency of a smooth function to resemble its tangent line over a sufficiently small interval.

A linear approximation uses a known value and local slope: $L(x)=f(a)+m(x-a)$ . The approximation error usually shrinks as $x$ moves closer to a for differentiable functions.

The point is not merely to reproduce a formula. A learner should be able to identify the quantities or geometric objects involved, explain why the relationship has its stated form, and recognize when the same idea appears in a graph, table, diagram, or model.

Textbook reading

Why the relationship works

Corners, cusps, discontinuities, and violent oscillation can defeat a single local line.

Textbook reading

A reliable way to work

Identify the anchor point and slope, write the line, compare exact and approximate values at several distances, and describe the error trend.

A graph appearing linear at one screen resolution is visual evidence, not a proof of differentiability.

After the symbolic work is complete, check the result. Depending on the lesson, this may mean substituting into an original equation, comparing coordinates, examining a graph, checking units, testing an interval, or confirming that every branch of a periodic solution has been included.

Textbook reading

What commonly goes wrong

A common error is using the secant slope from a large interval as though it were a local tangent slope.

The repair is to return to the definition and identify the first step where the invalid solution stops describing the original mathematical object. Later algebra cannot rescue a first step that changed the domain, orientation, branch, or meaning of the problem.

Textbook reading

Worked examples

Worked example 1

Zoom around $x=1$ on $y=x^2$ and build the tangent-line approximation from the predicted slope $2$ .

Solution

Begin by identifying the mathematical object and the information that fixes it. Identify the anchor point and slope, write the line, compare exact and approximate values at several distances, and describe the error trend. The relevant conditions are not optional bookkeeping: A graph appearing linear at one screen resolution is visual evidence, not a proof of differentiability. Following that structure gives $L(x)=1+2(x-1)=2x-1$ .

Why this works

Corners, cusps, discontinuities, and violent oscillation can defeat a single local line. The calculation and the representation should agree, so a graph, diagram, table, or substitution check should support the same conclusion.

Transfer example

Problem

Compare approximation errors at $x=1.1 \qquad 1.01$

Worked development

Identify the anchor point and slope, write the line, compare exact and approximate values at several distances, and describe the error trend. In this example, the first useful move is to make the defining structure visible rather than to search for a memorized answer. A linear approximation uses a known value and local slope: $L(x)=f(a)+m(x-a)$ . The approximation error usually shrinks as $x$ moves closer to a for differentiable functions. Then apply the conditions explicitly: A graph appearing linear at one screen resolution is visual evidence, not a proof of differentiability. Finish by checking the result in a second representation and explaining what the result means.

Interpretation

Local linearity supports estimation, measurement error analysis, and the conceptual meaning of derivatives.

Reasoning example

Problem

Analyze a corner at |x|.

Worked development

Identify the anchor point and slope, write the line, compare exact and approximate values at several distances, and describe the error trend. In this example, the first useful move is to make the defining structure visible rather than to search for a memorized answer. A linear approximation uses a known value and local slope: $L(x)=f(a)+m(x-a)$ . The approximation error usually shrinks as $x$ moves closer to a for differentiable functions. Then apply the conditions explicitly: A graph appearing linear at one screen resolution is visual evidence, not a proof of differentiability. Finish by checking the result in a second representation and explaining what the result means.

Interpretation

Local linearity supports estimation, measurement error analysis, and the conceptual meaning of derivatives.

Worked example 4: quick check

Use local linearity for sqrt(x) near $x=9$ with slope about $\frac{1}{6}$ to estimate $sqrt(9.3)$ .

Solution

Begin by identifying the mathematical object and the information that fixes it. Identify the anchor point and slope, write the line, compare exact and approximate values at several distances, and describe the error trend. The relevant conditions are not optional bookkeeping: A graph appearing linear at one screen resolution is visual evidence, not a proof of differentiability. Following that structure gives Approximately $3+\frac{0.3}{6}=3.05$ .

Why this works

Corners, cusps, discontinuities, and violent oscillation can defeat a single local line. The calculation and the representation should agree, so a graph, diagram, table, or substitution check should support the same conclusion.

Textbook reading

Application and interpretation

Local linearity supports estimation, measurement error analysis, and the conceptual meaning of derivatives.

A contextual answer must include units, a meaningful domain, and the assumptions that make the model plausible. An exact mathematical relationship should not be diluted into a decimal unless a measurement or comparison requires it.

Practice