Introduction

Type 1 ÷ 0 into a calculator and you get an error. Dividing by zero is a forbidden move in mathematics. So when learning limits and running into a situation where both numerator and denominator become zero, many people reflexively brush it off as, "ah, there is no answer."

But \(0 \div 0\) is a little different. It does not mean "there is no answer"; it means "this alone is not yet enough to know." In mathematics such a form is called an indeterminate form (a form that is not yet pinned down). And "not yet pinned down" also means it can be pinned down once we dig out more information.

The one sentence of this post is this. Even when the numerator and denominator both head toward zero, once you ask "how fast" each one heads there, the ratio settles on a single value. The move that reveals that "how fast" is cancelling. For a polynomial we cancel by dividing out a common factor; when a root is involved we cancel by multiplying by the conjugate. Let us follow the process step by step.

1. Why zero divided by zero is 'not yet known'

First let us pin down why \(0 \div 0\) fails to settle on a single value. Division \(a \div b\) asks, "how many times must we multiply \(b\) to get \(a\)?"

  • \(6 \div 2\) asks "how many times 2 makes 6" → exactly one answer, 3.
  • \(1 \div 0\) asks "how many times 0 makes 1" → 0 multiplied by anything is still 0, so there is no answer at all. Hence the error.
  • \(0 \div 0\) asks "how many times 0 makes 0" → any multiple of 0 is 0, so anything works as an answer. Hence it cannot be pinned to one value.

The third case is the key. \(0 \div 0\) is stuck not because it has no answer, but because it has too many candidates to pick just one. So "not yet known" is the accurate phrase, not "no answer."

Here is why this form shows up in limits. If, at some spot \(x = a\), the numerator and denominator become zero at the same time, that is usually a signal that both carry the same factor (a shared part that forces the value to zero). That common factor is exactly the culprit producing \(0 \div 0\), and it is what we clear away by cancelling.

  • Intuition: \(0 \div 0\) is a race asking "both runners arrived at the finish at zero — but who shrank faster?" Arrival alone cannot decide the winner; you have to watch the speed of approach.
  • Explanation: at a spot where the numerator and denominator both go to zero, a common factor is hiding.
  • Equation: cancel that common factor away, and you can plug the value into what remains to read off the limit.

2. Polynomials: cancel the common factor

Let us start with the cleanest case. Consider the following limit.

$$\lim_{x \to 3} \frac{x^2 - 9}{x - 3}$$

Plug in \(x = 3\) directly and the numerator becomes \(9 - 9 = 0\) and the denominator \(3 - 3 = 0\), giving \(\tfrac{0}{0}\), the "not yet known" form. Do not give up here. The fact that numerator and denominator both became zero is a signal that both carry the shared part \((x - 3)\). Indeed, factoring the numerator makes that part visible.

$$x^2 - 9 = (x + 3)(x - 3)$$

Now cancel the common factor \((x - 3)\) from numerator and denominator. This cancelling is valid only when \(x \ne 3\). But a limit looks only at the neighborhood of \(x = 3\), not at \(x = 3\) itself, so this condition does not get in the way of computing the limit at all.

$$\frac{x^2 - 9}{x - 3} = \frac{(x + 3)(x - 3)}{x - 3} = x + 3 \quad (x \ne 3)$$

What remains after cancelling is simply \(x + 3\). Plug \(x = 3\) into this and the value comes out with no obstacle.

$$\lim_{x \to 3} \frac{x^2 - 9}{x - 3} = \lim_{x \to 3} (x + 3) = 6$$

In the graph, the original expression \(\tfrac{x^2-9}{x-3}\) is exactly the same as the line \(y = x + 3\) except at the single point \(x = 3\). Since there is no value at that one point, the spot \((3, 6)\) is punched out as a hole. The value being approached (the limit value) stands perfectly well at \(6\); only the value landed on (the function value) is missing. The \(\tfrac{0}{0}\) that popped out when we plugged in was precisely a signal that we had stepped right onto this hole. Why the function value and the limit value can split like this was covered in detail in the previous post.

3. Expressions with a root: multiply by the conjugate to cancel

For polynomials the common factor showed up immediately through factoring. But when the numerator contains a root (a square root), the factor is not visible and it is hard to find something to cancel. The following limit is like that.

$$\lim_{x \to 3} \frac{\sqrt{x + 1} - 2}{x - 3}$$

Plug in \(x = 3\) and the numerator becomes \(\sqrt{4} - 2 = 0\) and the denominator \(0\), so it is \(\tfrac{0}{0}\) again. But there is no easy way to factor the numerator \(\sqrt{x+1} - 2\). The tool we use here is the conjugate (the partner expression). The conjugate of the root expression \(\sqrt{x+1} - 2\) is \(\sqrt{x+1} + 2\), with only the middle sign flipped.

Why multiply by exactly this? Because of the multiplication formula \((A - B)(A + B) = A^2 - B^2\): multiplying by the conjugate squares the root and makes it vanish cleanly. Since multiplying numerator and denominator by the same thing does not change the value, we multiply both top and bottom.

$$\frac{\sqrt{x+1} - 2}{x - 3} \cdot \frac{\sqrt{x+1} + 2}{\sqrt{x+1} + 2}$$

Tidy up the numerator first and the root gets squared away.

$$(\sqrt{x+1} - 2)(\sqrt{x+1} + 2) = (x + 1) - 4 = x - 3$$

Remarkably, the numerator has turned into \(x - 3\). Now the common factor to cancel against the denominator's \(x - 3\) has surfaced. Writing the whole thing out gives this.

$$\frac{x - 3}{(x - 3)(\sqrt{x+1} + 2)} = \frac{1}{\sqrt{x+1} + 2} \quad (x \ne 3)$$

Plug \(x = 3\) into what remains after cancelling, \(\tfrac{1}{\sqrt{x+1}\,+\,2}\), and again the value comes out with no obstacle.

$$\lim_{x \to 3} \frac{\sqrt{x+1} - 2}{x - 3} = \frac{1}{\sqrt{4} + 2} = \frac{1}{4}$$

Why the multiplication formula \((A-B)(A+B)=A^2-B^2\) holds, and why a root separates only across multiplication and division, was covered in the post on root separation laws. Multiplying by the conjugate is really using that same formula in reverse, in the "get rid of the root" direction.

4. Check it yourself with the interactive below

Both examples ran into the same wall on the surface, \(\tfrac{0}{0}\), yet once the move of cancelling cleared the wall away, each revealed a different value — \(6\) and \(\tfrac{1}{4}\). This is exactly what "\(0 \div 0\) does not settle on one value" looks like in practice: the same \(\tfrac{0}{0}\), yet different answers.

Below, switch modes and use the slider to bring \(x\) toward \(3\). You will see the numerator and denominator both shrink down to \(0\) side by side, while their ratio gathers steadily toward the target value (the green hole).

Clearing 0/0 away by cancelling
In polynomial mode the numerator and denominator of (x²−9)/(x−3) both go to 0 at x=3, yet the cancelled form x+3 gathers to 6. In conjugate mode, multiplying the root expression (√(x+1)−2)/(x−3) by the conjugate clears the root and the value gathers to one quarter. In both cases the graph is left with just a single hole.

In the panel the numerator and denominator each shrink toward \(0\) in red, but the value in the "numerator ÷ denominator" row does not waver — it sticks to the target. The numbers show directly that the ratio of the speeds at which the two head to zero is precisely the limit value.

5. Why this move is legitimate — a limit does not look at 'the point'

One doubt may arise here. "You said the cancelling is valid only when \(x \ne 3\), so is it really okay to use that cancelling to find the limit at \(x = 3\)?"

It is. And that reason is the heart of this whole post. A limit deliberately refuses to look at the point \(x = 3\) itself; it looks only at the neighbors on either side. What we seek is not "the value at \(x = 3\)" but "the value headed toward as \(x\) approaches \(3\)." The path of approach does not include \(x = 3\), so a cancelling that fails to be valid at that one point alone causes no trouble whatsoever for computing the limit.

To sum up, the order is always this.

  • Step 1 — Diagnose: if plugging in gives \(\tfrac{0}{0}\), read it not as "no answer" but as "a common factor to cancel is hiding here."
  • Step 2 — Move: for a polynomial, factor; when a root is involved, multiply by the conjugate — bring out that common factor and cancel it.
  • Step 3 — Substitute: plug the value into the clean expression that remains and read off the limit.

Closing

The most important attitude when you meet \(0 \div 0\) is to read it as a clue, not a dead end. The very fact that numerator and denominator became zero together is a map telling you, "a common factor to cancel is hiding here." Cancelling clears away that hidden part and reveals a limit value that plugging in alone could never show.

This view — that the contest is decided by the ratio of speeds — does not end here. The very definition of the derivative you will meet later is a story of resolving exactly this \(\tfrac{0}{0}\) form by cancelling, and L'Hôpital's rule, which goes further and differentiates numerator and denominator separately, is likewise an extension of the same intuition: "let us directly compare the speeds at which the two head to zero." This one wall we cleared today is, in a sense, the key that opens the door to the calculus ahead.