L'Hôpital's Rule: What It Is and When to Apply It

Try plugging in the number. Take $\lim_{x \to 0} \frac{\sin(x)}{x}$ and put $x = 0$ straight into it. You get $\frac{0}{0}$.

That's not a small number. It's not a big number. It's not really a number at all, it's a contradiction wearing a fraction's clothes. And yet this exact limit shows up constantly in calculus, and it has a perfectly clean answer: $1$.

So what do you do when direct substitution just breaks? You stop asking "what is it" and start asking "how fast is each side getting there." That shift is the entire idea behind L'Hôpital's rule.

Zero over zero isn't nothing, it's a tie

Picture two runners crossing a finish line at the exact same moment. Runner $f$ has a certain distance left to run. Runner $g$ has a certain distance left too. Right as they cross, both distances hit zero at once.

Now someone asks: "who was ahead, in terms of the ratio of their remaining distances?" At the instant they cross, that question sounds meaningless. Both numbers are zero. Zero divided by zero doesn't pick a winner.

But here's the thing. Just before the line, both runners were still moving, and each one has a speed. If you look at the last tiny sliver of the race instead of the exact instant of crossing, the ratio of how much ground they still had to cover stops being a coin flip. It settles down to the ratio of their speeds.

That's the whole trick, stated in plain English. Swap "speed" for "derivative," and "distance left" for "the function's value," and you've got L'Hôpital's rule.

Watching two functions race to zero

Let's make the runner's-finish-line picture concrete with real functions. Take $f(x) = \sin(x)$ and $g(x) = x$. Both are zero at $x = 0$, and as $x$ shrinks toward $0$, both $f(x)$ and $g(x)$ shrink toward $0$ right alongside it.

Drag $x$ down toward the finish line. Notice the bars for $f(x)$ and $g(x)$ both shrink together, and yet the ratio $f(x)/g(x)$ doesn't do anything scary. It calmly settles near $1$. Two quantities can both vanish and still leave behind a perfectly well-defined ratio. That's the entire mystery we need to explain, not "why is it zero," but "why does the ratio survive."

Why the ratio survives: zoom in and every curve goes straight

Here's the part that actually explains it, instead of just showing it. Take any smooth function and zoom in close enough around a point, and it stops looking curved. It starts looking like a straight line, its tangent line. That's basically the definition of a derivative: the slope of the line the function turns into once you zoom in far enough.

So if both $f$ and $g$ hit zero at the same point $c$, and you zoom in close enough, $f$ starts looking like the straight line $f'(c) \cdot (x - c)$, and $g$ starts looking like $g'(c) \cdot (x - c)$. Divide one by the other and the $(x - c)$ part, the leftover distance to the finish line, cancels clean out of both. What's left is just $\frac{f'(c)}{g'(c)}$.

Crank the zoom slider up. Watch the solid curves peel away from being curvy and lock onto their dashed tangent lines. The ratio of the two curves at the edge of the window locks onto $3$, which is exactly $f'(0)/g'(0)$. Nothing mysterious is happening here. You're just watching two curves turn into two straight lines through the origin, and the ratio of two lines through the origin is just the ratio of their slopes.

Writing the promise in symbols

Now that the idea is solid, the notation is just shorthand for it. If $\lim_{x \to c} f(x) = 0$ and $\lim_{x \to c} g(x) = 0$ (or both go to $\pm\infty$), and $f$ and $g$ are differentiable near $c$, then:

as long as that second limit actually exists, or heads to $\pm\infty$. And $c$ doesn't have to be a finite number. It can be $\infty$ too, the rule doesn't care, it only cares about the shape of the tie you're trying to break.

Let's use it. For $\lim_{x \to 0} \frac{\sin(x)}{x}$: differentiate top and bottom separately, giving $\frac{\cos(x)}{1}$. Plug in $x = 0$, and $\cos(0) = 1$. Limit found: $1$. No factoring tricks, no algebra gymnastics, just two derivatives and a substitution.

A zero on the bottom doesn't always mean "apply the rule"

Here's the misconception that trips people up more than anything else with this rule. Seeing a zero anywhere in a limit does not mean you get to differentiate. The rule only fires on a genuine tie: $\frac{0}{0}$ or $\frac{\infty}{\infty}$. Anything else, and differentiating top and bottom doesn't just fail to help, it gives you the wrong answer.

Click through the examples. Notice $\frac{x+3}{x+1}$ at $x \to 0$ plugs straight in to $\frac{3}{1}$, an ordinary number, no tie, no rule needed. And notice $\frac{x}{x-1}$ at $x \to 1$ gives $\frac{1}{0}$, which looks like it might qualify, but it doesn't. A nonzero number over zero isn't a tie, it's a blowup. The function is diverging, not sitting in limbo. Checking the form first, before you touch a single derivative, is the actual skill here.

Sometimes one derivative isn't enough

One more wrinkle. Differentiating top and bottom once doesn't guarantee you're done. If the new ratio $\frac{f'(x)}{g'(x)}$ is still $\frac{0}{0}$ or $\frac{\infty}{\infty}$ when you plug in $c$, you haven't broken the tie yet. You just have a smaller version of the same problem, so you apply the rule again, and keep going until the tie actually breaks.

Take $\lim_{x \to 0} \frac{x - \sin(x)}{x^3}$. Plug in $0$ and you get $\frac{0}{0}$. Differentiate once: $\frac{1 - \cos(x)}{3x^2}$, plug in $0$, still $\frac{0}{0}$. Differentiate again: $\frac{\sin(x)}{6x}$, plug in $0$, still $\frac{0}{0}$. Only on the third round does it break.

Watch the bars stay flat at zero for rounds $0$, $1$, and $2$, both racers still perfectly tied. Then round $3$ finally shows daylight between them: $\cos(0) = 1$ on top, a plain constant $6$ on the bottom. Limit: $\frac{1}{6}$. The rule doesn't promise one derivative fixes everything. It promises that if you keep differentiating through genuine ties, you'll eventually reach one that isn't.

The short version

A limit that plugs in to $\frac{0}{0}$ or $\frac{\infty}{\infty}$ isn't broken. It's a photo finish, both sides reaching zero (or infinity) at the exact same moment, so comparing their raw values tells you nothing. L'Hôpital's rule breaks the tie by comparing rates instead: differentiate the top, differentiate the bottom, and check the new ratio.

Before you reach for it, check the form. Only a real $\frac{0}{0}$ or $\frac{\infty}{\infty}$ qualifies, anything else already has an answer sitting right there. And if differentiating once still leaves you tied, differentiate again. The tie always breaks eventually, you just have to keep racing until it does.