From Secant Slopes to a Derivative at a Point
A secant line passes through two points on a function’s graph. If one point is fixed at \(x=a\) and the other is at \(x\), the slope of that secant line is a quotient involving the function values at those inputs. As \(x\) approaches \(a\), the second point approaches the fixed point.
If these secant slopes approach a finite number, that number describes the graph’s instantaneous slope at \(x=a\). The x-to-a form of the difference quotient expresses this idea directly: compare \(f(x)\) with the fixed value \(f(a)\), divide by the change in input \(x-a\), and take the limit as \(x\) approaches \(a\).
The condition \(x\ne a\) matters because the denominator would be zero at \(x=a\). A two-sided derivative at an interior point requires the quotient to approach the same finite value from both sides. If the function is defined only on one side of a point, a one-sided derivative may instead be considered; here, we focus on the usual two-sided definition.
Why the Quotient Is a Secant Slope
For \(x\ne a\), the graph of \(f\) has points \((a,f(a))\) and \((x,f(x))\). The slope formula for the line through two points is the change in vertical coordinate divided by the change in horizontal coordinate. Substituting these points gives exactly the x-to-a difference quotient.
Thus the limit in the definition asks whether the slopes of these secant lines settle toward one finite value as the moving input \(x\) gets closer to \(a\). When they do, that limiting value is the derivative at \(a\), and it gives the tangent slope in the limit-of-secants interpretation.
The numerator and denominator both depend on how far \(x\) is from \(a\). Their ratio measures the average change in \(f\) per unit change in input between the two inputs. The limit captures what that ratio approaches as the interval between the inputs shrinks toward zero.
This definition is not the value of the quotient at \(x=a\). It is the value approached by the quotient for nearby inputs different from \(a\). That distinction allows the derivative to exist even though the expression inside the limit is undefined at the point itself.
Worked Example: A Positive-Slope Linear Function
Worked Example: A Positive-Slope Linear Function
Find the derivative of \(f(x)=3x-2\) at \(a=4\) from the x-to-a definition.
First evaluate the function at the fixed input:
Use the definition, with \(f(4)=10\):
The cancellation is valid for \(x\ne4\), which is precisely where the difference quotient is defined. After simplifying, the quotient has the constant value \(3\) for every nearby \(x\ne4\), so its limit is \(3\).
Check: For any \(x\ne4\), the secant slope from \((4,10)\) to \((x,3x-2)\) is \(\frac{(3x-2)-10}{x-4}=\frac{3(x-4)}{x-4}=3\). This agrees with the limit.
A Complete Justification for the Linear Case
The cancellation in the example is not a lucky simplification. For any linear function \(f(x)=mx+b\), the difference quotient at an input \(a\) simplifies to a constant, provided \(x\ne a\). Here \(m\) and \(b\) are fixed real numbers.
For \(f(x)=mx+b\), substituting \(a\) gives \(f(a)=ma+b\).
Subtract the fixed value from \(f(x)\), then divide by \(x-a\): \(\frac{(mx+b)-(ma+b)}{x-a}\).
The constants cancel and the numerator factors as \(m(x-a)\), so the quotient equals \(m\).
Since the quotient is \(m\) for every \(x\ne a\), its limit as \(x\to a\) is \(m\).
Written as a derivation, each equality follows directly from the definition of the linear function and the condition \(x\ne a\):
This argument establishes the result for every real \(m\), \(b\), and \(a\). It also explains why the slope of a linear graph is its derivative at each point: the secant slope is already the same constant \(m\) at every pair of distinct inputs, so taking a limit does not change it.
Worked Example: A Negative-Slope Linear Function
Worked Example: A Negative-Slope Linear Function
Find the derivative of \(g(x)=-2x+5\) at \(a=1\) using the limit definition.
Evaluate the function at \(1\):
Substitute into the x-to-a form and simplify:
The quotient equals \(-2\) for every \(x\ne1\), so the limit is \(-2\).
Check: The fixed point is \((1,3)\), and the moving point is \((x,-2x+5)\). Their secant slope is \(\frac{(-2x+5)-3}{x-1}=\frac{-2(x-1)}{x-1}=-2\) whenever \(x\ne1\), confirming the result.
Worked Example: A Fractional Slope
Worked Example: A Fractional Slope
Find the derivative of \(h(x)=\frac{1}{2}x+4\) at \(a=-2\).
First calculate the fixed function value:
Now form and simplify the quotient:
The factor \(x+2\) cancels for \(x\ne-2\). The simplified quotient is constant, so the limit is \(\frac12\).
Check: The fixed point is \((-2,3)\). For any \(x\ne-2\), the secant slope is \(\frac{(\frac12x+4)-3}{x-(-2)}=\frac{\frac12(x+2)}{x+2}=\frac12\), as required.
A Common Mistake: Substituting Before Taking the Limit
In the x-to-a quotient, putting \(x=a\) directly makes both the numerator and denominator zero: \(f(a)-f(a)=0\) and \(a-a=0\). That does not evaluate the limit. It only shows that the original quotient is undefined at the point where the limit is being taken.
The correct order is to form the quotient, simplify it on its domain near \(a\), and then take the limit. A cancellation involving \(x-a\) is allowed because the limit concerns inputs \(x\ne a\) approaching \(a\), not the value at \(a\).
How to Use the Definition Reliably
For a derivative at one specified input, keep the fixed point \(a\) visible throughout the calculation. The expression \(f(a)\) is a single fixed number once \(a\) has been chosen, whereas \(f(x)\) varies as \(x\) approaches \(a\). Mixing those roles is a common source of errors.
- Calculate \(f(a)\) carefully before substituting it into the quotient.
- Write the denominator as \(x-a\), preserving the order used in the numerator.
- Simplify the quotient only where \(x\ne a\).
- Take the limit after simplification, and state the resulting derivative value.
For linear functions, the subtraction produces a factor \(x-a\), which cancels with the denominator. For other functions, the algebra may require different steps, and the limit may or may not exist. The definition itself remains the same: the derivative exists when the difference quotient approaches a finite value.
The x-to-a form is especially useful when a particular point is given explicitly. It places the fixed input directly in the expression, so the two points on the graph are visibly \((a,f(a))\) and \((x,f(x))\). The limit then describes what happens to the secant slope as the second point approaches the first.
Check Your Understanding
For each question, use the x-to-a definition rather than a slope shortcut.
- For \(f(x)=4x+1\), write the difference quotient for the derivative at \(a=2\), and simplify it for \(x\ne2\).
- For \(g(x)=-3x+2\), calculate \(g(-1)\) and then find \(g'(-1)\) from the limit definition.
- Why is the quotient \(\frac{f(x)-f(a)}{x-a}\) not evaluated by substituting \(x=a\) directly?
- In the definition, what geometric quantity does the quotient represent when \(x\ne a\)?
- For a linear function \(f(x)=mx+b\), why does its x-to-a quotient have the constant value \(m\) whenever \(x\ne a\)?
