Why Polynomial Zero-Over-Zero Limits Need Care
Some limits of polynomial quotients cannot be evaluated by direct substitution: both the numerator and denominator become zero. The expression \(0/0\) is not a value for the limit. It signals that substitution alone has not determined the answer.
L'Hôpital's Rule can resolve this situation by comparing the derivatives of the numerator and denominator. For polynomials, differentiation is especially manageable. Sometimes one differentiation is enough; sometimes both derivatives still vanish at the endpoint, so another application is needed. The key is to check the rule's conditions at every stage rather than differentiate automatically.
When L'Hôpital's Rule Applies
Here we consider a finite endpoint \(a\) and polynomials \(P\) and \(Q\). Direct substitution gives the indeterminate form \(0/0\) precisely when \(P(a)=0\) and \(Q(a)=0\). That fact alone does not justify applying L'Hôpital's Rule: the quotient must also be defined near \(a\), and the derivative quotient must have a limit.
Polynomials are differentiable everywhere. A nonzero polynomial has only finitely many roots, so if \(Q\) is not the zero polynomial, a root at \(a\) is isolated: \(Q(x)\ne 0\) in a sufficiently small deleted interval around \(a\). The same observation applies to \(Q'\) when it is not the zero polynomial. These facts help check the local denominator conditions, but they do not replace checking that each successive derivative quotient is still of type \(0/0\) when another application is proposed.
For example, if the first derivatives satisfy \(P'(a)=Q'(a)=0\), the first application has not yet produced a finite numerical ratio by substitution. A second application may be justified if its own conditions hold. If instead \(P'(a)/Q'(a)\) is already defined and gives the derivative quotient's limit, stop there.
Why the Rule Works
The reason for the one-application rule comes from the Cauchy Mean Value Theorem, which relates the changes in two differentiable functions to their derivatives at a point between the endpoints. For polynomials, the continuity and differentiability required by that theorem are automatic.
Take \(x\ne a\) close enough to \(a\) that \(Q(x)\ne 0\) and \(Q'\) is nonzero between \(a\) and \(x\).
On the interval with endpoints \(a\) and \(x\), there is a point \(c\) between them such that \(\frac{P(x)-P(a)}{Q(x)-Q(a)}=\frac{P'(c)}{Q'(c)}\).
Since \(P(a)=Q(a)=0\) and \(Q(x)\ne 0\), the equality becomes \(\frac{P(x)}{Q(x)}=\frac{P'(c)}{Q'(c)}\).
Because \(c\) lies between \(a\) and \(x\), \(c\) approaches \(a\) as \(x\) does. If \(P'(t)/Q'(t)\) tends to \(L\) as \(t\to a\), then \(P'(c)/Q'(c)\) tends to \(L\), so \(P(x)/Q(x)\) tends to \(L\) as well.
This proves the stated finite-limit case. Repeated use is simply a sequence of applications of this result: each application requires its own zero values at \(a\), local denominator conditions, and a derivative quotient with a limit. The proof does not say that a quotient and the quotient of its derivatives are equal everywhere.
One Application
Worked Example: A Linear Denominator
Evaluate \(\lim_{x\to 2}\frac{x^2-4}{x-2}\).
First substitute \(x=2\) into the numerator and denominator:
The form is \(0/0\). Both expressions are polynomials, and the denominator \(x-2\) is nonzero for \(x\ne 2\). Its derivative is \(1\), which is nonzero nearby. Differentiate numerator and denominator:
The derivative quotient has limit \(\lim_{x\to 2}\frac{2x}{1}=4\). L'Hôpital's Rule therefore gives:
Check: the derivative quotient has value \(2(2)/1=4\) at the endpoint, and the original numerator and denominator both vanish there. The result and the rule's conditions agree.
Repeated Differentiation
After one application, check the new numerator and denominator at the same endpoint. If both are still zero, a further application may be appropriate. For polynomials this can happen when the original expressions vanish to higher order: early derivatives may also vanish at the endpoint, while later ones do not. Each extra differentiation must be justified by the conditions of the rule, not merely by the fact that the previous step produced \(0/0\).
Worked Example: Two Applications
Evaluate \(\lim_{x\to 1}\frac{x^4-2x^2+1}{x^3-3x+2}\).
Substitution gives \(1-2+1=0\) in the numerator and \(1-3+2=0\) in the denominator. Both are polynomials. The denominator is not the zero polynomial, so its root at \(1\) is isolated. Differentiate once:
At \(x=1\), these derivative values are \(4-4=0\) and \(3-3=0\). Thus the first derivative quotient still gives \(0/0\); it does not yet give a value by substitution. The polynomial \(Q'(x)=3x^2-3\) is nonzero for \(x\ne 1\) sufficiently close to \(1\), so another application can be considered. Differentiate again:
Now \(P''(1)=12-4=8\) and \(Q''(1)=6\). The second derivative quotient has limit \(8/6=4/3\). The successive applications therefore give:
Check: the two first derivatives both vanish at \(1\), while the second derivative values are \(8\) and \(6\), whose ratio is \(4/3\). The denominator and its first derivative are nonzero in a sufficiently small deleted interval around \(1\), as required at the respective stages.
When Several Derivatives Are Needed
The same logic can require a third application. The important bookkeeping is to keep the derivative sequence visible and test the endpoint values at each stage. The next example uses polynomials whose first and second derivatives both vanish at the endpoint.
Worked Example: Three Applications
Evaluate \(\lim_{x\to 1}\frac{x^3-3x^2+3x-1}{2x^3-6x^2+6x-2}\).
At \(x=1\), the numerator is \(1-3+3-1=0\), and the denominator is \(2-6+6-2=0\). Both are nonzero polynomials. Their roots at \(1\) are isolated, so the original quotient is defined in a sufficiently small deleted interval around \(1\). The first derivatives are:
At \(1\), these give \(3-6+3=0\) and \(6-12+6=0\). The first derivative quotient is still \(0/0\). The second derivatives are:
At \(1\), these give \(6-6=0\) and \(12-12=0\). The second derivative quotient is also \(0/0\). The third derivatives are:
The third derivative quotient is the constant \(6/12=1/2\). The first and second denominator derivatives are nonzero in a sufficiently small deleted interval around \(1\): each is a nonzero polynomial with an isolated root there. Applying the rule at each stage gives:
Check: the endpoint values at the first two derivative stages are \(0/0\), and the third-stage values are \(6/12=1/2\). This confirms both the stopping point and the final arithmetic.
A Common Mistake: Continuing Without Checking
A common error is to treat repeated differentiation as an automatic procedure. Consider \(\lim_{x\to 1}\frac{x^2-1}{x-1}\). Substitution gives \(0/0\), and one application gives the derivative quotient \(\frac{2x}{1}\), whose limit is \(2\). It is incorrect to apply the rule again just because the original quotient began as \(0/0\). The new numerator and denominator are \(2x\) and \(1\), and at \(x=1\) their values are \(2\) and \(1\), not \(0\) and \(0\).
The incorrect extra step would differentiate again and write \(\lim_{x\to 1}\frac{2}{0}\), which is not a justified application of L'Hôpital's Rule. The correction is to stop after the first application: the derivative quotient already has limit \(2\), so the original limit is \(2\). The condition \(0/0\) must be checked anew before every application.
Another distinction matters: L'Hôpital's Rule is a theorem about limits, not a formula for the derivative of a quotient. In general, the derivative of \(P(x)/Q(x)\) is \(\big(P'(x)Q(x)-P(x)Q'(x)\big)/Q(x)^2\), where \(Q(x)\ne 0\); it is not \(P'(x)/Q'(x)\). The derivative ratio is used to evaluate a particular limit when the rule's hypotheses are met.
A Reliable Procedure
For a polynomial quotient at a finite endpoint, organize the work in this order:
- Substitute the endpoint into the original numerator and denominator. Confirm that both values are zero.
- Check that the original denominator is nonzero in a deleted neighborhood of the endpoint.
- Differentiate numerator and denominator separately, then evaluate both derivatives at the endpoint.
- If the derivative quotient has a limit, apply the rule and stop. If both derivative values are zero, check the conditions again before differentiating once more.
- State the resulting limit and verify the final derivative values and ratio.
This process keeps the distinction between an indeterminate form and a limit visible. A zero-over-zero substitution is a reason to investigate, not a conclusion about the limit. The derivatives provide the comparison only when the hypotheses justify it.
Check Your Understanding
For each question, check endpoint values and rule conditions before deciding how many differentiations are justified.
- For \(\lim_{x\to 3}\frac{x^2-9}{x-3}\), what form appears on direct substitution, and what limit follows from L'Hôpital's Rule?
- Why is it necessary to check the numerator and denominator again before a second application?
- For \(\lim_{x\to 1}\frac{x^3-3x+2}{x^2-1}\), what are the numerator and denominator values after one differentiation at \(x=1\)?
- State the role of the Cauchy Mean Value Theorem in justifying L'Hôpital's Rule.
- Why is the equality \( (P/Q)'=P'/Q'\) not a valid general differentiation rule?
