The Central Idea
For two distinct points on a graph, the slope between them is an average rate of change. Geometrically, that slope belongs to the line joining the two points.
That line is called a secant line.
If we move the second point closer and closer to the first, the secant line can approach a limiting position. When that limiting line exists, it is the tangent line.
Its slope represents the function's local rate of change at the point.
\[ \boxed{ \text{secant slope} \longrightarrow \text{tangent slope} } \]This transition from an interval to a point is one of the central ideas of calculus.
Review: Average Rate of Change
For a function \(y=f(x)\), the average rate of change from \(x=a\) to \(x=b\) is
\[ \boxed{ \frac{f(b)-f(a)}{b-a} }, \qquad a\ne b. \]The numerator measures the change in output:
\[ \Delta y=f(b)-f(a), \]while the denominator measures the change in input:
\[ \Delta x=b-a. \]Therefore,
\[ \boxed{ \text{average rate of change} = \frac{\Delta y}{\Delta x} }. \]What Is a Secant Line?
A secant line is a line passing through two distinct points on a curve.
If the points are
\[ P=(a,f(a)) \]and
\[ Q=(b,f(b)), \]then the slope of the secant line through \(P\) and \(Q\) is
\[ \boxed{ m_{\text{sec}} = \frac{f(b)-f(a)}{b-a} }. \]This is exactly the average rate of change of \(f\) on the interval from \(a\) to \(b\).
Secant Slope and Average Rate Are the Same Quantity
The formulas
\[ \frac{\Delta y}{\Delta x} \]and
\[ \frac{f(b)-f(a)}{b-a} \]can be interpreted in two different but equivalent ways.
- As a rate, they measure average output change per unit input change.
- As a slope, they measure the steepness of the secant line.
Algebra and geometry are describing the same relationship.
Example: A Secant Line on a Parabola
Consider
\[ f(x)=x^2. \]Use the points corresponding to \(x=1\) and \(x=3\).
\[ f(1)=1, \qquad f(3)=9. \]The points are
\[ P=(1,1), \qquad Q=(3,9). \]The secant slope is
\[ m_{\text{sec}} = \frac{9-1}{3-1} = \frac{8}{2} = \boxed{4}. \]Thus, the function changes by an average of \(4\) output units for each input unit across this interval.
The Secant Line Equation
Once the secant slope is known, the secant line itself can be written using point-slope form.
In the previous example,
\[ m=4 \]and the line passes through \((1,1)\). Therefore,
\[ y-1=4(x-1). \]Simplifying,
\[ \boxed{y=4x-3}. \]This line crosses the parabola at the two selected points.
A Secant Line Depends on Two Points
A single point does not determine a unique secant line. We need two distinct points.
Fix
\[ P=(a,f(a)). \]Then choose another point
\[ Q=(b,f(b)). \]Different choices of \(b\) generally produce different secant slopes.
This is especially important for nonlinear functions, whose rates change from interval to interval.
From an Interval Toward a Point
Suppose we want to understand the behavior of a curve at
\[ x=a. \]We cannot compute
\[ \frac{f(a)-f(a)}{a-a}, \]because this becomes
\[ \frac{0}{0}, \]which is undefined.
Instead, we keep a second point distinct from the first and move it progressively closer.
Using a Nearby Point
Let the fixed point be
\[ P=(a,f(a)). \]Instead of writing the second input as \(b\), write it as
\[ a+h, \]where \(h\ne0\).
Then the nearby point is
\[ Q=(a+h,f(a+h)). \]The horizontal change is
\[ \Delta x=(a+h)-a=h. \]The vertical change is
\[ \Delta y=f(a+h)-f(a). \]The Nearby Secant Slope
The slope through
\[ (a,f(a)) \]and
\[ (a+h,f(a+h)) \]is
\[ \boxed{ \frac{f(a+h)-f(a)}{h} }, \qquad h\ne0. \]This expression is a difference quotient.
For now, its geometric meaning is most important: it is the slope of a secant line through a fixed point and a nearby point.
What Happens as \(h\) Gets Smaller?
When \(h\) is large, the second point may be far from the first.
When \(h\) is small,
\[ a+h \]lies close to \(a\), so the second point lies close to the fixed point on the graph.
We can examine the secant slopes for values such as
\[ h=1,\quad0.5,\quad0.1,\quad0.01, \]and ask whether those slopes appear to approach a particular number.
Example: Secants Approaching a Tangent
Consider
\[ f(x)=x^2 \]near
\[ x=2. \]The fixed point is
\[ P=(2,4). \]A nearby point has input \(2+h\), so the secant slope is
\[ \frac{f(2+h)-f(2)}{h}. \]Since
\[ f(2+h)=(2+h)^2, \]we obtain
\[ \frac{(2+h)^2-4}{h}. \]Simplifying the Secant Slope
Expand:
\[ (2+h)^2=4+4h+h^2. \]Then
\[ \frac{(2+h)^2-4}{h} = \frac{4+4h+h^2-4}{h}. \]Simplify the numerator:
\[ \frac{4h+h^2}{h}. \]Factor \(h\):
\[ \frac{h(4+h)}{h}. \]Because the secant calculation uses \(h\ne0\), we may cancel:
\[ \boxed{4+h}. \]A Table of Nearby Secant Slopes
For \(f(x)=x^2\) near \(x=2\), the secant slope is
\[ 4+h. \]| \(h\) | Nearby input \(2+h\) | Secant slope \(4+h\) |
|---|---|---|
| \(1\) | \(3\) | \(5\) |
| \(0.5\) | \(2.5\) | \(4.5\) |
| \(0.1\) | \(2.1\) | \(4.1\) |
| \(0.01\) | \(2.01\) | \(4.01\) |
As \(h\) becomes smaller and positive, the secant slopes approach \(4\).
Approaching from the Other Side
We should also examine nearby points with inputs less than \(2\). These correspond to negative values of \(h\).
| \(h\) | Nearby input \(2+h\) | Secant slope \(4+h\) |
|---|---|---|
| \(-1\) | \(1\) | \(3\) |
| \(-0.5\) | \(1.5\) | \(3.5\) |
| \(-0.1\) | \(1.9\) | \(3.9\) |
| \(-0.01\) | \(1.99\) | \(3.99\) |
From this side, the slopes also approach \(4\).
The Tangent Line
A tangent line describes the local direction of a curve at a point.
For \(f(x)=x^2\) at \(x=2\), the nearby secant slopes approach
\[ 4. \]Therefore, the tangent line at
\[ (2,4) \]has slope
\[ \boxed{4}. \]Equation of the Tangent Line
The tangent line passes through
\[ (2,4) \]and has slope \(4\).
Using point-slope form,
\[ y-4=4(x-2). \]Simplifying,
\[ \boxed{y=4x-4}. \]Near \(x=2\), this line follows the direction of the parabola closely.
What Is Local Rate of Change?
A local rate of change describes how rapidly a function is changing near a particular input.
When a tangent slope exists, that slope gives the local rate.
\[ \boxed{ \text{local rate at }x=a = \text{slope of the tangent line at }x=a }. \]This differs from average rate, which compares two distinct inputs across an interval.
Average Rate versus Local Rate
| Average Rate | Local Rate |
|---|---|
| Uses two distinct input values | Describes behavior at one input |
| Associated with a secant line | Associated with a tangent line |
| Summarizes an interval | Describes nearby behavior |
| \(\dfrac{f(b)-f(a)}{b-a}\) | Obtained from limiting nearby secant slopes |
The local rate is not found by simply setting \(b=a\) in the average-rate formula. Instead, we study what happens as \(b\) approaches \(a\).
The Idea of a Limit
A limit describes the value a quantity approaches.
In the parabola example, the nearby secant slope was
\[ 4+h. \]As \(h\) approaches \(0\),
\[ 4+h \]approaches
\[ 4. \]Symbolically,
\[ \lim_{h\to0}(4+h)=4. \]At this stage, the important idea is approaching a value, not a formal theory of limits.
Why We Do Not Substitute \(h=0\) Too Early
The original secant quotient is
\[ \frac{f(a+h)-f(a)}{h}. \]Direct substitution of \(h=0\) gives
\[ \frac{f(a)-f(a)}{0} = \frac{0}{0}, \]which is undefined.
We first simplify the expression for nonzero \(h\), and then study what the simplified values approach as \(h\) becomes arbitrarily small.
Derivative Notation as a Preview
When the local rate of change of \(f\) exists at \(x=a\), it is commonly written
\[ \boxed{f'(a)}. \]This is read as
"\(f\) prime of \(a\)."
Geometrically,
\[ f'(a) = \text{slope of the tangent line at }x=a. \]Later lessons will develop systematic methods for finding derivatives.
The Derivative as a Local Rate
The derivative gives a compact mathematical name to the local rate.
Conceptually,
\[ \boxed{ f'(a) = \text{the value approached by nearby secant slopes} }. \]A formal notation for this idea is
\[ f'(a) = \lim_{h\to0} \frac{f(a+h)-f(a)}{h}, \]when that limit exists.
For now, treat this formula as a map of the idea. The next lessons will unpack its algebra and computation.
Linear Functions Are a Special Case
Consider
\[ f(x)=3x+2. \]Any two distinct points on this line have slope
\[ 3. \]Therefore, every secant line has slope \(3\), and the tangent line is simply the line itself.
\[ \boxed{ \text{average rate} = \text{local rate} = 3 }. \]This happens because a linear function has constant rate of change.
Nonlinear Functions Have Changing Local Rates
For a nonlinear function, different parts of the graph can have different steepness and direction.
For
\[ f(x)=x^2, \]the curve is not equally steep everywhere.
Near negative \(x\)-values, tangent slopes are negative. At the vertex, the tangent is horizontal. For positive \(x\)-values, tangent slopes are positive and become steeper as \(x\) increases.
Local rate captures this changing behavior.
Positive, Negative, and Zero Tangent Slopes
The sign of a tangent slope gives immediate information about local behavior.
- If the tangent slope is positive, the function is locally increasing.
- If the tangent slope is negative, the function is locally decreasing.
- If the tangent slope is zero, the tangent line is horizontal.
A horizontal tangent often appears at a local maximum or minimum, although a horizontal tangent alone does not guarantee an extremum.
Horizontal Tangents
A horizontal line has slope
\[ 0. \]Therefore, if a smooth curve has a horizontal tangent at \(x=a\), then its local rate there is
\[ \boxed{f'(a)=0}. \]Such points will become important when we study critical points, extrema, and optimization.
A Tangent Line Is Not Merely a Line That Touches Once
In elementary geometry, a tangent to a circle is often introduced as a line that touches the circle at exactly one point.
For general curves, that description is too restrictive.
A tangent line is better understood through local direction or as the limiting position of nearby secant lines.
A tangent line may cross a curve and still be the tangent line at the point of contact.
Example: A Tangent Can Cross the Curve
Consider
\[ f(x)=x^3 \]at
\[ x=0. \]The graph passes through the origin and has a horizontal tangent there:
\[ y=0. \]Yet the curve crosses this tangent line as \(x\) changes sign.
Therefore, "touches but does not cross" is not a reliable general definition of tangent.
Local Linear Approximation
A smooth curve often looks increasingly like its tangent line when viewed very close to the point of tangency.
If the tangent slope at \(x=a\) is \(m\), then near \(a\),
\[ f(x) \approx f(a)+m(x-a). \]This is a local linear approximation.
The tangent line does not replace the curve globally, but it can approximate the curve well over a sufficiently small interval.
Example: Tangent-Line Estimation
For
\[ f(x)=x^2 \]at \(x=2\), we found
\[ f(2)=4 \]and tangent slope
\[ m=4. \]The tangent line is
\[ y=4+4(x-2). \]To estimate \(f(2.01)\),
\[ f(2.01) \approx 4+4(0.01) = 4.04. \]The exact value is
\[ (2.01)^2=4.0401, \]so the local linear estimate is very close.
Local Rate in Motion
Suppose \(s(t)\) gives the position of an object at time \(t\).
Over a time interval from \(t_1\) to \(t_2\), the average velocity is
\[ \frac{s(t_2)-s(t_1)}{t_2-t_1}. \]This is the slope of a secant line on the position-versus-time graph.
The local rate at one instant is the object's instantaneous velocity.
Geometrically, instantaneous velocity is the tangent slope on the position graph.
Units of Local Rate
Local rate has the same type of units as average rate:
\[ \boxed{ \frac{\text{output units}}{\text{input units}} }. \]If \(s(t)\) is measured in meters and \(t\) in seconds, then both average and local rates have units
\[ \frac{\text{meters}}{\text{second}}. \]If \(C(q)\) is measured in dollars and \(q\) in units produced, then a local rate has units of dollars per unit.
Estimating a Local Rate from a Table
Sometimes we have data rather than a formula.
Suppose a function has the following values near \(x=5\):
| \(x\) | \(f(x)\) |
|---|---|
| \(4.9\) | \(11.71\) |
| \(5.0\) | \(12.00\) |
| \(5.1\) | \(12.31\) |
The left-hand secant slope is
\[ \frac{12.00-11.71}{5.0-4.9} = \frac{0.29}{0.1} = 2.9. \]The right-hand secant slope is
\[ \frac{12.31-12.00}{5.1-5.0} = \frac{0.31}{0.1} = 3.1. \]These values suggest a local rate near
\[ \boxed{3}. \]Using a Symmetric Secant Estimate
Another useful estimate uses points equally spaced on opposite sides of the target input.
With the same data,
\[ \frac{f(5.1)-f(4.9)}{5.1-4.9} = \frac{12.31-11.71}{0.2} = \frac{0.60}{0.2} = \boxed{3}. \]This centered average can give a strong estimate of the local slope when nearby data are reasonably smooth.
Estimating Tangent Slope from a Graph
When only a graph is available, we can estimate the tangent line visually.
- Identify the point of interest.
- Draw or imagine the line that best matches the curve's local direction.
- Choose two convenient points on that tangent line.
- Compute rise over run using those two line points.
Because the tangent itself may be estimated, the resulting local rate should usually be reported as an approximation.
Zooming In on a Smooth Curve
One way to think about differentiability is through magnification.
Near many points on a smooth curve, repeated zooming makes the graph appear straighter and straighter.
The line that emerges from this local view is the tangent line.
This is why the tangent line provides a natural linear model for nearby behavior.
Not Every Point Has an Ordinary Tangent Slope
Some graphs have points where a unique finite tangent slope does not exist.
Examples can include:
- sharp corners;
- cusps;
- certain discontinuities;
- vertical tangents, whose ordinary slope is not finite.
Thus, the existence of a function value does not automatically guarantee the existence of an ordinary derivative at that point.
Example: A Corner
Consider
\[ f(x)=|x|. \]At the origin, the graph has a sharp corner.
Secant slopes approaching from the right behave like
\[ 1, \]while those approaching from the left behave like
\[ -1. \]Because the two sides do not approach the same slope, there is no single tangent slope at the origin in the ordinary derivative sense.
Secants Can Reveal Local Shape
Nearby secant slopes tell us more than just one numerical rate.
As we move through a curve, changes in secant and tangent slopes can reveal:
- where the function is increasing or decreasing;
- where the graph becomes steeper or flatter;
- where horizontal tangents may occur;
- where local maxima or minima may occur;
- how the graph bends.
Later lessons will organize these observations using derivatives, critical points, and concavity.
Secant Slopes Across Smaller Intervals
Suppose we study a function at \(x=a\) using a sequence of intervals:
\[ [a,a+1], \] \[ [a,a+0.1], \] \[ [a,a+0.01], \] \[ [a,a+0.001]. \]Each interval has its own average rate and corresponding secant slope.
If those slopes stabilize near one value, that value is evidence for the local rate at \(a\).
Local Does Not Mean Exactly One Point of Data
A local rate belongs to a point, but we learn about it by examining values arbitrarily close to that point.
This distinction is important.
The local rate is not obtained from an isolated function value such as
\[ f(a). \]It depends on how the function behaves around \(a\).
Common Error: Using Only the Function Value
Suppose
\[ f(2)=7. \]This tells us the point
\[ (2,7) \]lies on the graph.
It does not tell us the tangent slope there.
Many different curves can pass through the same point while having different local rates.
Common Error: Setting the Two Inputs Equal
The secant slope formula requires distinct inputs:
\[ \frac{f(b)-f(a)}{b-a}, \qquad b\ne a. \]Replacing \(b\) directly with \(a\) produces division by zero.
The correct idea is not to make the points equal during the secant calculation, but to study what happens as the second point approaches the first.
Common Error: Confusing a Small Interval with a Zero Interval
A very small nonzero value such as
\[ h=0.0001 \]is still not zero.
The secant quotient is defined for that value if the corresponding function values are defined.
The local-rate idea depends on being able to make \(h\) as small as needed without requiring \(h=0\) inside the original quotient.
Common Error: Assuming Every Tangent Is Horizontal
A tangent line can have any allowable local direction.
It may have:
- positive slope;
- negative slope;
- zero slope;
- or, in special cases, a vertical direction.
Horizontal tangents are important, but they are only one possibility.
Common Error: Assuming the Tangent Cannot Cross the Curve
The tangent line is determined by local direction, not by a rule that it must remain on one side of the graph.
Some tangent lines cross their curves.
Therefore, the general calculus meaning of tangent should be connected to local approximation and limiting secant lines rather than merely "touching."
Common Error: Treating an Estimated Local Rate as Exact
If a local rate is estimated from a graph or a small collection of data, the result usually carries approximation error.
For example,
\[ f'(5)\approx3 \]is different from asserting
\[ f'(5)=3. \]Use the approximation symbol when the available information supports only an estimate.
A Reliable Procedure for Finding a Secant Slope
- Identify the two distinct input values.
- Find the corresponding function values.
- Compute the output change using final minus initial.
- Compute the input change in the same order.
- Divide the output change by the input change.
- Interpret the sign and units.
- Recognize the result as both an average rate and a secant slope.
A Reliable Procedure for Estimating a Local Rate
- Choose the input \(x=a\) where the local rate is desired.
- Keep the point \((a,f(a))\) fixed.
- Choose nearby inputs on one or both sides of \(a\).
- Compute secant slopes using those nearby points.
- Move the nearby inputs progressively closer to \(a\).
- Look for a value toward which the secant slopes stabilize.
- Interpret that value as the tangent slope when the pattern supports a unique limit.
- Use appropriate units and approximation notation when necessary.
Example: Full Secant-to-Tangent Workflow
Let
\[ f(x)=x^2+x \]and investigate the local rate at
\[ x=1. \]First,
\[ f(1)=1^2+1=2. \]Use a nearby input \(1+h\):
\[ f(1+h) = (1+h)^2+(1+h). \]Expand:
\[ f(1+h) = 1+2h+h^2+1+h = 2+3h+h^2. \]Form the secant quotient:
\[ \frac{f(1+h)-f(1)}{h} = \frac{2+3h+h^2-2}{h}. \]Simplify:
\[ \frac{3h+h^2}{h} = 3+h. \]As \(h\) approaches \(0\),
\[ 3+h\to3. \]Therefore, the local rate is
\[ \boxed{3}. \]Writing the Tangent Line in the Full Example
For
\[ f(x)=x^2+x \]at \(x=1\), we found the point
\[ (1,2) \]and local slope
\[ 3. \]Point-slope form gives
\[ y-2=3(x-1). \]Therefore,
\[ \boxed{y=3x-1}. \]Practice 1: Secant Slope
Let
\[ f(x)=x^2+1. \]Find the secant slope from \(x=1\) to \(x=4\).
Practice 2: Secant Line Equation
For
\[ f(x)=x^2, \]find the equation of the secant line through the points corresponding to \(x=0\) and \(x=2\).
Practice 3: Nearby Secant
Let
\[ f(x)=x^2 \]and use the fixed input \(x=3\). Find the secant slope from \(x=3\) to \(x=3.1\).
Practice 4: Difference Quotient
For
\[ f(x)=x^2 \]at \(x=3\), simplify
\[ \frac{f(3+h)-f(3)}{h}, \qquad h\ne0. \]Practice 5: Estimate the Local Rate
Using your result from Practice 4, determine the value approached by the secant slope as \(h\) approaches \(0\).
Practice 6: Tangent Line
For
\[ f(x)=x^2 \]at \(x=3\), use the local rate from Practice 5 to find the tangent-line equation.
Practice 7: Linear Function
Let
\[ f(x)=-4x+7. \]What are the average rate on any nonzero interval and the local rate at any point?
Practice 8: Local Rate from Data
A function has the nearby values
| \(x\) | \(f(x)\) |
|---|---|
| \(1.9\) | \(7.61\) |
| \(2.0\) | \(8.00\) |
| \(2.1\) | \(8.41\) |
Use the points at \(x=1.9\) and \(x=2.1\) to estimate the local rate at \(x=2\).
Practice 9: Interpret the Sign
A function has a tangent slope of
\[ -2.5 \]at \(x=7\). What does the sign tell you about the function's local behavior?
Practice 10: Identify the Error
A student wants the local rate of \(f\) at \(x=4\) and writes
\[ \frac{f(4)-f(4)}{4-4}=0. \]Explain the error and describe the correct idea.
Practice Answers
Practice 1
First compute
\[ f(1)=2, \qquad f(4)=17. \]Then
\[ \frac{17-2}{4-1} = \frac{15}{3} = \boxed{5}. \]Practice 2
The two points are
\[ (0,0) \qquad\text{and}\qquad (2,4). \]The slope is
\[ \frac{4-0}{2-0}=2. \]Since the line passes through the origin,
\[ \boxed{y=2x}. \]Practice 3
\[ f(3)=9, \qquad f(3.1)=9.61. \]Therefore,
\[ \frac{9.61-9}{3.1-3} = \frac{0.61}{0.1} = \boxed{6.1}. \]Practice 4
\[ \frac{(3+h)^2-9}{h} = \frac{9+6h+h^2-9}{h} \] \[ = \frac{6h+h^2}{h} = \boxed{6+h}. \]Practice 5
As \(h\to0\),
\[ 6+h\to6. \]Therefore, the local rate is
\[ \boxed{6}. \]Practice 6
The point is
\[ (3,9) \]and the slope is \(6\). Thus,
\[ y-9=6(x-3). \]Therefore,
\[ \boxed{y=6x-9}. \]Practice 7
A linear function has constant slope equal to its coefficient of \(x\). Therefore,
\[ \boxed{-4} \]is both its average rate on every nonzero interval and its local rate at every point.
Practice 8
\[ \frac{8.41-7.61}{2.1-1.9} = \frac{0.80}{0.2} = \boxed{4}. \]So the local rate at \(x=2\) is estimated as approximately \(4\).
Practice 9
The negative slope indicates that the function is locally decreasing near \(x=7\). Its output decreases at about \(2.5\) output units per input unit locally, with units determined by the quantities involved.
Practice 10
The expression
\[ \frac{f(4)-f(4)}{4-4} \]is not \(0\); it is
\[ \frac{0}{0}, \]which is undefined.
To investigate the local rate, use a nearby distinct input such as \(4+h\), compute
\[ \frac{f(4+h)-f(4)}{h}, \qquad h\ne0, \]and then study what the secant slopes approach as \(h\to0\).
What to Remember
The slope of the secant line through
\[ (a,f(a)) \]and
\[ (b,f(b)) \]is
\[ \boxed{ \frac{f(b)-f(a)}{b-a} }. \]This is also the average rate of change.
Using a nearby point \(a+h\), the secant slope becomes
\[ \boxed{ \frac{f(a+h)-f(a)}{h} }, \qquad h\ne0. \]- secant lines describe average change across intervals;
- tangent lines describe local direction at a point;
- nearby secant slopes can approach a tangent slope;
- the tangent slope is the local rate of change;
- \(f'(a)\) is standard notation for the derivative or local rate at \(a\);
- a local rate depends on nearby behavior, not merely on the function value;
- not every graph has a unique finite tangent slope at every point.
Where We Go Next
The secant-to-tangent idea gives a geometric picture of local change. To use that idea efficiently, we need to understand the algebraic expressions that measure change between nearby inputs and the discrete patterns that appear when inputs are separated by fixed steps.
The next lesson, Difference Quotients and Finite Differences, develops those tools and connects algebraic change, tabular patterns, secant slopes, and the emerging derivative concept.