Why Repeated Applications Need Care
A single application of L'Hôpital's Rule may leave a limit unresolved. For example, differentiating the numerator and denominator can produce another zero-over-zero or infinity-over-infinity form. In that situation, another application may be justified—but only after checking the rule's conditions again.
The key is to treat every application as a new step with its own eligibility check. Repeated differentiation is not an instruction to keep differentiating until an answer appears. Stop as soon as the quotient has a limit that can be evaluated directly, or as soon as it no longer has one of the forms for which the rule applies.
The Rule at Each Stage
For a limit as \(x\) approaches \(a\), suppose \(f\) and \(g\) are differentiable on a punctured interval around \(a\), \(g(x)\ne 0\) there, and \(g'(x)\ne 0\) there. If \(f(x)\) and \(g(x)\) both tend to zero, or both tend to infinity in magnitude, and the derivative quotient has a limit, L'Hôpital's Rule allows the original quotient to be evaluated using that derivative quotient. The limit may be finite or infinite.
For repeated applications, write the successive numerator and denominator pairs as \(f_0,g_0\), then \(f_1=f_0'\), \(g_1=g_0'\), and so on. At stage \(k\), the rule can be applied to \(f_k/g_k\) only if that pair has a qualifying indeterminate form and the rule's hypotheses hold for it. In particular, the derivative of the current denominator must be nonzero in a punctured interval where the quotient is being considered.
Each arrow represents a justified use of the rule, not merely a formal differentiation. The sequence ends when a quotient has a determined limit or when its form is no longer \(0/0\) or \(\infty/\infty\). A changed form is not a reason to apply L'Hôpital again; it is a signal to evaluate the limit by other direct reasoning.
Why the Rule Works in the Zero-Over-Zero Case
A short justification explains why differentiating can reveal the original quotient's limit. Suppose \(f(x)\to 0\) and \(g(x)\to 0\) as \(x\to a\), and the functions satisfy the differentiability conditions of L'Hôpital's Rule. Define \(f(a)=0\) and \(g(a)=0\); these definitions make the functions continuous at \(a\). For \(x\ne a\), the Cauchy Mean Value Theorem applied between \(x\) and \(a\) gives a point \(c\) between them such that
Since \(f(a)=g(a)=0\), the left side is \(f(x)/g(x)\). As \(x\to a\), the point \(c\), which lies between \(x\) and \(a\), also approaches \(a\). Therefore, if \(f'(x)/g'(x)\to L\), the equality shows that \(f(x)/g(x)\to L\). This explains the basic mechanism: near a common zero, the quotient of the function changes is represented by a quotient of derivatives at an intermediate point.
This argument justifies one application. To repeat it, the new numerator and denominator must themselves satisfy the relevant conditions; the justification does not automatically license unlimited differentiation.
Worked Examples
Worked Example: Repeating the Rule Twice at Zero
Evaluate \(\displaystyle \lim_{x\to 0}\frac{1-\cos x}{x^2}\).
Substitution gives \(1-\cos 0=0\) and \(0^2=0\), so the form is \(0/0\). The numerator and denominator are differentiable near zero, and the derivative of the denominator, \(2x\), is nonzero for \(x\ne 0\) near zero. Apply the rule once:
The new quotient is still \(0/0\), since \(\sin 0=0\) and \(2(0)=0\). Its denominator has derivative \(2\), which is nonzero. Apply the rule again:
Check: The second application differentiates \(\sin x\) to \(\cos x\) and \(2x\) to \(2\). The resulting quotient has a direct finite limit, so there is no reason to differentiate again. The positive result is also consistent with \(1-\cos x\ge 0\) near zero and \(x^2>0\) for \(x\ne 0\).
Worked Example: Stop When the Form Changes
Evaluate \(\displaystyle \lim_{x\to\infty}\frac{x^2}{e^x}\).
As \(x\to\infty\), both \(x^2\) and \(e^x\) tend to infinity, so the original form is \(\infty/\infty\). The denominator's derivative \(e^x\) is nonzero, and both functions are differentiable. Apply L'Hôpital's Rule:
The new quotient is again \(\infty/\infty\). The denominator's derivative is still \(e^x\ne 0\), so a second application is allowed:
At this stage the numerator tends to \(2\) and the denominator tends to infinity. This is not \(0/0\) or \(\infty/\infty\); the quotient already tends to zero. Stop here.
Check: Both applications used a denominator whose derivative is \(e^x\), which never vanishes. The final quotient \(2/e^x\) tends to zero because its numerator is constant and its denominator grows without bound. The original quotient is positive, consistent with the zero limit.
Worked Example: Stop When the Limit Is Directly Determined
Evaluate \(\displaystyle \lim_{x\to\infty}\frac{x^2+3x}{x^2+x}\).
The numerator and denominator both tend to infinity, so the starting form is \(\infty/\infty\). Both are differentiable, and the derivative of the denominator, \(2x+1\), is nonzero for \(x\) sufficiently large. Apply L'Hôpital's Rule:
The new numerator and denominator still tend to infinity, and the derivative of the new denominator is \(2\ne 0\). Apply the rule once more:
The last quotient is already constant, so the limit is determined and the process stops.
Check: Divide the original numerator and denominator by \(x^2\) to verify the result independently:
This agrees with the repeated-application result.
Common Mistake: Applying the Rule After Eligibility Ends
Consider the second example. After two applications, the quotient is \(2/e^x\), which tends to zero. It would be incorrect to say, “Apply L'Hôpital again: the derivative of \(2\) is \(0\), so the answer is \(0/e^x\).” That is not a valid application: \(2/e^x\) is not \(0/0\) or \(\infty/\infty\). Its limit is already determined directly.
Another error is to keep differentiating simply because the previous step used the rule. The number of applications is not fixed in advance. One limit may require a second application, another may be resolved after one, and another may cease to qualify after differentiation.
A Practical Stopping Checklist
At each stage, use this sequence of checks before differentiating again:
Find the limit of the current numerator and the current denominator separately. Confirm that the form is \(0/0\) or \(\infty/\infty\).
Confirm differentiability near the limiting value and that the current denominator's derivative is nonzero on the punctured interval being used.
Differentiate the current numerator and denominator separately. Do not use the quotient rule on their ratio as a substitute for L'Hôpital's Rule.
If the new quotient is eligible and unresolved, another application may be possible. If its limit is determined or its form is not eligible, stop applying the rule.
Why the Stopping Rule Matters
Repeated applications are useful when each derivative quotient remains indeterminate. They can expose a limit that was hidden in the original expression, as in the first example. But the theorem controls what each application proves: it relates the limit of one eligible quotient to the limit of its derivative quotient. It does not say that differentiation preserves limits in general.
For a reliable solution, show the form before each application, show the derivatives that produce the next quotient, and state why the process stops. This makes clear not only how the answer was obtained, but also why every step is permitted.
Check Your Understanding
For each question, pay attention to the form at the current stage rather than assuming that the rule can always be repeated.
- In \(\displaystyle \lim_{x\to 0}\frac{\sin x-x}{x^3}\), what form results after one application, and does it permit another application?
- State two conditions that must be checked again before applying the rule to a new derivative quotient.
- After obtaining \(2/e^x\) in the second worked example, why is another application unnecessary and not eligible?
- In the third worked example, why does the second application finish the calculation?
- Explain in one or two sentences why repeated use of L'Hôpital's Rule is not permission to differentiate indefinitely.
