The Tangent Slope As A Limit

From Nearby Secants to a Tangent Slope

A secant line passes through two points on a function's graph. Its slope measures the average change in the function's output between the two corresponding inputs. If one point is held fixed while the other moves closer to it, the secant slopes may approach a single value. That limiting value is the tangent slope at the fixed point, when the limit exists and is finite.

The word limit matters. A secant slope uses two distinct inputs, so it is not itself the slope at one point. The tangent slope is the value that nearby secant slopes approach—not necessarily a value attained by any particular secant.

In the limit, \(x\) approaches \(a\) but is not equal to \(a\). At \(x=a\), the displayed quotient would have denominator zero, so the secant slope is defined only for a second input distinct from the fixed one. The limit describes what happens to the quotient for nearby inputs other than \(a\).

More precisely, saying that the secant slopes approach \(L\) means that for every \(\varepsilon>0\), there is a \(\delta>0\) such that whenever \(0<|x-a|<\delta\), the secant slope differs from \(L\) by less than \(\varepsilon\). The condition \(0<|x-a|\) keeps the second input distinct from \(a\). Requiring inputs on both sides of \(a\) gives a two-sided tangent slope; the left- and right-hand limiting values must agree.

Worked Example: The Tangent Slope of a Squared Function

This calculation also gives a direct justification of the limit. The difference between the secant slope and \(4\) is \(\left|(x+2)-4\right|=|x-2|\). Given any \(\varepsilon>0\), choose \(\delta=\varepsilon\). Then \(0<|x-2|<\delta\) implies \(\left|(x+2)-4\right|=|x-2|<\varepsilon\). This verifies that the limiting value is \(4\), not merely that a few sampled slopes look close to it.

What the Limit Does—and Does Not—Say

The limit concerns a whole pattern of secant slopes as the second input approaches the fixed input. It does not claim that every secant slope equals the tangent slope. In the squared-function example, the secant from input \(2\) to input \(3\) has slope \(5\), while the limiting tangent slope is \(4\). That secant uses a second point one unit away; it is just one member of the family being considered.

The value \(L\) is uniquely determined if the limit exists. A quantity cannot approach two different numbers in the precise sense of the limit definition: if two proposed limits differed, choose \(\varepsilon\) smaller than half their distance. Sufficiently close secant slopes would then have to lie within \(\varepsilon\) of both numbers, which is impossible. Thus the definition cannot assign two tangent slopes at the same point.

The requirement that the limit exist is essential. If the secant slopes from the left approach one value and those from the right approach a different value, then there is no single two-sided limiting value, so this definition gives no two-sided tangent slope there. It is not enough that the quotient be defined for nearby inputs, or that some selected secants have slopes that look similar.

Worked Example: A Reciprocal Function

A Justification for Using the Nearby Secants

The definition is motivated by the geometric role of secants. A secant uses the fixed graph point \((a,f(a))\) and a second graph point \((x,f(x))\). The change in vertical coordinate is \(f(x)-f(a)\); the change in horizontal coordinate is \(x-a\). The slope is vertical change divided by horizontal change. As the second point approaches the fixed point, the corresponding secant slopes describe the local limiting direction of the graph. If those slopes settle to one finite value, that value is the tangent slope.

The definition can be justified without treating any one secant as the tangent. It asks whether all sufficiently nearby secant slopes are as close as desired to the same number. In precise terms, the proof has two parts: identify a candidate value from the quotient, then show that the quotient's distance from that candidate becomes arbitrarily small when \(x\) is sufficiently close to \(a\). The squared-function example did this directly: its quotient differs from \(4\) by exactly \(|x-2|\). The closeness of the inputs therefore controls the closeness of the slopes.

Worked Example: Rationalizing a Secant Quotient

Common Pitfall: Substituting One Secant Slope

A common error is to calculate a secant slope for a convenient nearby point and call that number the tangent slope. For \(f(x)=x^2\) at \(a=2\), taking \(x=3\) gives

The arithmetic is correct, but the conclusion “the tangent slope is \(5\)” is not. It identifies only the slope of the secant through the inputs \(2\) and \(3\). The correction is to keep the second input variable, simplify the quotient, and then take its limit:

Therefore the tangent slope is \(4\). The single secant slope \(5\) is not wrong as a secant slope; it is wrong only if it is presented as the limiting tangent slope.

Why the Limiting Slope Matters

Secant slopes describe changes over intervals, while a tangent slope captures the value approached as that interval shrinks toward one point. This distinction provides a precise way to move from average change over a nonzero interval to local change at a point. It is useful in interpreting graphs and in later work with rates of change, but the essential idea here is the limit itself.

Whenever calculating a tangent slope by this definition, keep track of three details:

• The fixed input \(a\) and its value \(f(a)\) must be used consistently.
• The quotient is formed only for \(x\ne a\); simplify it while preserving that restriction.
• The final answer is a limiting value, which requires the nearby secant slopes to approach one finite number from both sides.