L'Hôpital's Rule

The most common mistake with l'hopital's rule is reaching for it the moment a limit looks hard. That is wrong. L'Hôpital's Rule is a narrow tool for one specific situation: a limit that collapses into the indeterminate forms 0/0 or ∞/∞ when you plug in the point. If the form is anything else, the rule is not just useless, it is actively misleading. You do not use it because a fraction has a variable in it. You use it because the numerator and denominator both head to zero, or both head to infinity, and no other technique is faster. Before you differentiate anything, check the form. That check is the entire game.

When the form is right, the rule is brutally simple: differentiate the top, differentiate the bottom, and take the limit again. You are not differentiating the quotient as a quotient, and you are not using the product rule on the whole fraction. You take the derivative of the numerator and the derivative of the denominator separately, then re-evaluate. If the result is still 0/0 or ∞/∞, you do it again. The rule lets you repeat as long as the conditions hold. But the conditions are precise, and skipping them is how students burn a page of work on a limit that was never indeterminate to begin with.

L'Hôpital's Rule states that if f and g are differentiable near a point a (except possibly at a itself), if g′(x) is not zero near a (except possibly at a), and if the limit of f(x)/g(x) as x approaches a is an indeterminate form 0/0 or ∞/∞, then the limit of f(x)/g(x) equals the limit of f′(x)/g′(x), provided that second limit exists or is infinite. The theorem is attributed to Johann Bernoulli, but the name stuck, and so did the rule's restrictions.

Three conditions matter in practice. First, the limit must be exactly 0/0 or ∞/∞. A form like 0/∞ is not indeterminate; it is simply zero, and L'Hôpital's Rule does not apply. Second, the functions must be differentiable in an interval around the point, which rules out corners and cusps. Third, the limit of the derivative quotient must exist. If f′/g′ oscillates or blows up without settling, the rule fails, even if the original limit exists. That last condition is the one textbooks mention in passing and students forget until a problem punishes them.

The Classic and the Workhorse

The classic 0/0 case is lim_x→0 sin x / x = 1. Direct substitution gives 0/0, so the rule applies. Differentiate the top to get cos x, differentiate the bottom to get 1, and the new limit is cos 0 / 1 = 1. That is the whole procedure in one line. But the case that teaches you more is lim_x→0 (1 − cos x) / x². Substitution gives 0/0. Differentiate once to get sin x / (2x), which is still 0/0. Differentiate again to get cos x / 2, which evaluates to 1/2. Two applications, one answer, and a clear view of why repeating works.

When Algebra Beats the Rule

Here is what goes wrong when you skip the check. Suppose a student sees 0/0 and immediately differentiates, but the denominator has a factor that cancels. Differentiating is not wrong, but it is slower than factoring x out of the numerator and denominator first. L'Hôpital's Rule is a hammer, and not every 0/0 is a nail. If the function is a rational polynomial with a hole, factor it. If it has a radical, rationalize. The rule is for when algebra has run out of moves, not for when you cannot be bothered to factor.

A Clean Case and a Trap

For the ∞/∞ form, the canonical case is lim_x→∞ e^x / x² = ∞. Substitution gives ∞/∞, so the rule applies. Differentiate to get e^x / (2x), still ∞/∞. Differentiate again to get e^x / 2, which goes to ∞. That case is clean because exponential growth outruns any polynomial, and the rule reveals it in two steps. But the ∞/∞ form hides a trap: the rule can loop forever without resolving the limit, and that is a signal to stop differentiating.

Consider lim_x→∞ √(x² + 1) / x. Substitution gives ∞/∞. Differentiate the top to get x / √(x² + 1), differentiate the bottom to get 1, and the new limit is x / √(x² + 1), which is again ∞/∞. Differentiate again and you are back where you started, chasing your tail. The correct move is to divide numerator and denominator by x, giving √(1 + 1/x²) / 1, which approaches 1. L'Hôpital's Rule is not a panacea, and the moment the derivative quotient stops simplifying, you switch to algebra. The rule is a tool, not a ritual.

When Repeated Application Works

Repeated application is allowed as long as each new quotient still satisfies the conditions. The condition to recheck is the indeterminate form: after one differentiation, if the new limit is still 0/0 or ∞/∞, you can differentiate again. The case lim_x→0 (e^x − 1 − x) / x² requires two passes. First pass gives (e^x − 1) / (2x), still 0/0. Second pass gives e^x / 2, which evaluates to 1/2. That is the standard pattern, and it works because each differentiation simplifies the expression.

When Repeated Application Fails

The failure case is when repeated application makes things worse. If you differentiate a quotient and the new limit oscillates, or if the derivative quotient has no limit, the rule stops applying. The textbook caution is blunt: L'Hôpital's Rule may fail if the limit of f′/g′ does not exist, even when the original limit exists. An example is lim_x→∞ (x + sin x) / x. Substitution gives ∞/∞. Differentiate to get (1 + cos x) / 1, which oscillates between 0 and 2 and has no limit. The original limit is 1, found by dividing through by x. The rule fails, and the algebra succeeds. That is why you always have an escape plan.

How to Handle 0·∞

The forms 0·∞ and ∞−∞ are indeterminate, but L'Hôpital's Rule does not touch them directly. You must convert them into 0/0 or ∞/∞ first. For 0·∞, rewrite the product as a fraction. The classic case is lim_x→0⁺ x ln x = 0. Write it as ln x / (1/x), which is ∞/∞ after substitution. Differentiate to get (1/x) / (−1/x²) = −x, which approaches 0. That conversion is the entire trick, and it works because moving one factor to the denominator changes the form.

How to Handle ∞−∞

For ∞−∞, you combine terms over a common denominator. The case lim_x→0 (1/sin x − 1/x) = 0 starts as ∞−∞. Combine to get (x − sin x) / (x sin x), which is 0/0. Apply the rule once to get (1 − cos x) / (sin x + x cos x), still 0/0. Apply again to get sin x / (2 cos x − x sin x), which evaluates to 0/2 = 0. The conversion is mandatory; skipping it and applying the rule to the original difference is a category error. The rule only works on a quotient, so make the quotient first.

L'Hôpital's Rule fails in three concrete ways, and you need to recognize each one. First, the form is not 0/0 or ∞/∞, such as 0/∞. That form is not indeterminate; the limit is simply 0, and applying the rule gives a wrong answer. Second, the derivative quotient has no limit, even when the original limit exists, as with the (x + sin x)/x case. Third, the rule loops without progress, as with √(x² + 1)/x, where each differentiation returns a similar form.

When the rule loops, stop and switch techniques. Factor, rationalize, or divide by the highest power. For limits at infinity, dividing by the largest term in the denominator almost always works. For limits near zero, use known Taylor expansions or the squeeze theorem, but never force the rule. A limit that resists two applications of L'Hôpital's Rule is a signal that the problem wants a different approach. The rule is not a badge of honor; it is a shortcut, and shortcuts expire.

Condition	What to Check	If It Fails
Indeterminate form	Limit is exactly 0/0 or ∞/∞	Do not use the rule; simplify algebraically
Differentiability	f and g differentiable near the point, except possibly at it	Check for corners, cusps, or discontinuities
Denominator not zero	g′(x) ≠ 0 near the point, except possibly at it	Rule does not apply; find another method
Derivative limit exists	lim f′/g′ exists or is infinite	Original limit may still exist; use algebra
Repeated application	Each new quotient is still 0/0 or ∞/∞	Stop after two tries; switch techniques

Shining Cases and Failing Cases

The lhospital rule cases that stick are the ones where the rule shines and the ones where it fails. The shining cases are lim_x→0 sin x / x = 1, lim_x→∞ e^x / x² = ∞, and lim_x→0⁺ x ln x = 0. The failing cases are lim_x→∞ (x + sin x)/x = 1 and lim_x→∞ √(x² + 1)/x = 1, both of which require algebra. Memorize the pattern, not the problems: if the derivative quotient simplifies, keep going; if it does not, stop.

When to Use It and When to Avoid It

When to use l'hopital's rule is the question that separates a passing student from a stuck one. Use it when direct substitution yields 0/0 or ∞/∞, when the functions are differentiable, and when you have already tried factoring or rationalizing. Do not use it when the form is 0·∞ or ∞−∞ unless you convert first. Do not use it when the derivative quotient oscillates. Do not use it when a one-sided limit is involved and the function is not differentiable from that side. The rule is a scalpel, not a chainsaw.

When to use l'hopital's rule comes down to a three-step decision. Step one: plug in the point. If you get a number, the limit is that number, and you are done. Step two: if you get 0/0 or ∞/∞, ask if the expression simplifies. If it factors, factor it. If it has a radical, rationalize. Only if those fail, or if the algebra is uglier than a derivative, apply the rule. Step three: after one application, re-evaluate. If the form persists, apply again, but cap it at two or three passes before switching to another method.

The failure case is when a limit involves a piecewise function or an absolute value. For example, lim_x→0 |x| / x does not exist because the left-hand limit is −1 and the right-hand limit is 1. L'Hôpital's Rule is useless here because the function is not differentiable at 0, and the one-sided limits differ. Check the domain before you differentiate. If the function has a corner at the point, the rule is void. That is a named failure mode, and it is the one that costs exam points.

Forms Needing a Logarithm

Indeterminate forms like 0⁰, 1^∞, and ∞⁰ require a different conversion, one that uses the exponential. For example, lim_x→0⁺ x^x = 1. Write it as e^ln(xx) = e^{x ln x}, then evaluate the exponent as a 0·∞ form, which converts to 0/0. The exponent goes to 0, so the limit is e⁰ = 1. Similarly, lim_x→∞ x^1/x = 1, and lim_x→0 (1+x)^1/x = e. These are not directly L'Hôpital problems; they are L'Hôpital problems after a logarithm.

The Rule as One Move in a Larger Playbook

The takeaway is that the rule is one move in a larger playbook. Indeterminate forms tell you that direct substitution has failed, but they do not tell you which tool to use. Sometimes the tool is factoring, sometimes it is the squeeze theorem, and sometimes it is L'Hôpital's Rule. The mistake is assuming the form dictates the tool. It does not. The form only tells you that more work is needed. Do that work, and the rule becomes a natural part of your process rather than a reflex.