Limits are the backbone of Calculus. They describe what happens to a function as it gets incredibly close to a specific point, even if it never actually touches it. This concept allows us to handle infinity and division by zero.
Why does calculus need them? Because both derivatives and integrals are secretly limits: a derivative is the limit of average slopes over shrinking intervals, and an integral is the limit of rectangle sums as the rectangles get thinner. Master limits and the rest of calculus stops feeling like magic.
When Substitution Fails
Usually, you can just plug the number in. Lim x->2 of (x + 3) is just 5. But what about (x² - 1)/(x - 1) as x->1? Plugging in 1 gives 0/0. This is an indeterminate form. It doesn't mean "undefined"; it means "we have more work to do".
L'Hôpital's Rule
A superpower for solving limits. If you get 0/0 or ∞/∞, you can take the derivative of the top and bottom separately, then try again.
Worked Example: Two Ways to Crack 0/0
Let's evaluate lim x→1 of (x² − 1)/(x − 1), the classic indeterminate form.
Method 1 — Factoring:
- Factor the numerator: x² − 1 = (x − 1)(x + 1).
- Cancel the common factor: (x − 1)(x + 1)/(x − 1) = x + 1, valid for every x except exactly 1.
- Now substitute: as x → 1, the expression approaches 1 + 1 = 2.
Method 2 — L'Hôpital's Rule:
- Confirm the form is 0/0 at x = 1. It is — this check is mandatory.
- Differentiate top and bottom separately: derivative of x² − 1 is 2x; derivative of x − 1 is 1.
- Take the new limit: 2x/1 → 2·1 = 2.
Both roads give the same answer, which is a great way to double-check yourself on an exam. The function has a "hole" at x = 1 — it's never actually defined there — yet the limit tells us exactly what value the hole surrounds.
Limits at Infinity
Limits also describe long-run behavior. As x grows enormous, 1/x shrinks toward 0 — so lim x→∞ of 1/x = 0. For ratios of polynomials, only the highest powers matter in the long run: (3x² + 5x)/(2x² + 7) approaches 3/2, because the lower-order terms become irrelevant next to the x² giants. This is exactly how horizontal asymptotes are computed, and it's the tool economists and physicists use to ask "what does this system settle into eventually?"
Continuity
A function is continuous if you can draw it without lifting your pen. Mathematically, this means the limit from the left equals the limit from the right, which equals the function's value at that point. If any of the three pieces disagree — the function jumps, has a hole, or blows up to infinity — the function is discontinuous there.
Common Mistakes to Avoid
- Treating 0/0 as 0 or as undefined. An indeterminate form means the limit could be anything — 0, 7, infinity — until you do more algebra. Never stop at 0/0.
- Using L'Hôpital without checking the form. The rule ONLY applies to 0/0 or ∞/∞. Applying it to a limit like 5/0 or 3/2 produces confident nonsense.
- Differentiating with the quotient rule. L'Hôpital says differentiate the numerator and denominator separately — not differentiate the whole fraction as a quotient. Mixing these up is extremely common.
- Ignoring one-sided limits. For functions with jumps or absolute values, the left and right limits can differ. If they disagree, the two-sided limit does not exist, no matter what substitution suggests.
Frequently Asked Questions
Does a limit existing mean the function is defined at that point?
No — and that's the whole point. The limit describes the function's behavior near a point, not at it. A function can have a hole at x = 1 while its limit there is a perfectly clean number like 2.
What's the difference between a limit "not existing" and being "infinite"?
If a function grows without bound (like 1/x² near 0), we often write the limit as ∞ to describe that behavior — but strictly, the limit does not exist as a real number. If the left and right sides approach different values, the limit doesn't exist at all, not even in the infinite sense.
Can I always use L'Hôpital's rule instead of factoring?
Usually, but not always wisely. Some limits require applying the rule several times, and a few (like x/√(x²+1) as x→∞) loop forever under repeated differentiation. Factoring, conjugates, or dividing by the highest power is often faster and less error-prone.