When Substitution Produces Zero Over Zero
For a polynomial quotient, direct substitution is often an efficient first step. If the denominator has a nonzero value at the point, the quotient law allows us to evaluate the limit by evaluating the numerator and denominator separately. But sometimes substitution gives zero in both places. That result is a signal to pause: it does not give the limit.
The notation \(0/0\) describes the pair of substitution results, not a number that can be used as the quotient’s limit. The quotient is undefined at the point if its denominator is zero there, but a limit concerns values near the point. Those nearby values may approach a finite number, grow without bound, or fail to approach any one number.
The Quotient Law and Its Condition
The quotient law requires more than a denominator that is nonzero at the point when we simply inspect function values. It requires the relevant limits to exist and the denominator’s limit to be nonzero. In particular, if \(\lim_{x\to a} f(x)=L\) and \(\lim_{x\to a} g(x)=M\ne 0\), then the quotient limit exists and is \(L/M\).
When the functions are polynomials, their limits equal their values at \(a\). Thus, for a polynomial quotient \(P(x)/Q(x)\), if \(Q(a)\ne 0\), substitution gives the limit \(P(a)/Q(a)\). If instead \(P(a)=Q(a)=0\), the quotient law’s nonzero-denominator condition fails. That does not prove that the limit fails to exist; it means this particular law cannot determine the answer.
There is a second caution: values at a point alone do not establish limits for arbitrary functions. A function can be assigned any value at \(a\) while behaving differently nearby. The quotient law concerns limits as \(x\) approaches \(a\), not just the values \(f(a)\) and \(g(a)\).
Why Zero Over Zero Is Inconclusive
To see why the same substitution result can lead to different outcomes, compare quotients whose numerator and denominator both tend to zero. In each case, algebra reveals what the quotient does at nearby inputs. One quotient approaches a finite number; another grows without bound; a third does not approach a limit at all. Therefore, the substitution result \(0/0\) cannot by itself distinguish these possibilities.
Worked Example: A Finite Limit After Simplification
Evaluate \(\lim_{x\to 3}\frac{x^2-9}{x-3}\), if it exists.
First substitute \(x=3\) into the numerator and denominator:
The substitution result is \(0/0\), so it does not determine the limit. Rewrite the numerator using the identity \(x^2-9=(x-3)(x+3)\). For \(x\ne 3\), the quotient is therefore
This equality is valid for \(x\ne 3\), which is exactly the condition relevant to approaching 3 through nearby inputs. Hence
Check the value using the simplified expression: when \(x\) is close to 3, its difference from 6 is \(\lvert(x+3)-6\rvert=\lvert x-3\rvert\), which can be made arbitrarily small by taking \(x\) close enough to 3. Thus the simplified expression does approach 6, confirming the result. The original quotient is undefined at \(x=3\), but that does not prevent its limit from existing.
Worked Example: Zero Over Zero Can Lead to Unbounded Values
Examine \(\lim_{x\to 0}\frac{x^2}{x^4}\).
Substitution gives \(0^2=0\) in the numerator and \(0^4=0\) in the denominator. Again, \(0/0\) is inconclusive. For every \(x\ne 0\), simplify the quotient:
As \(x\) approaches 0 through nonzero values, \(x^2\) is positive and approaches 0, so \(1/x^2\) grows without bound. More explicitly, let \(M>0\). If \(0<\lvert x\rvert<1/\sqrt{M}\), then \(x^2<1/M\), and therefore \(1/x^2>M\). This verifies that the quotient increases beyond every positive bound as \(x\to 0\); it does not have a finite real limit.
Check the simplification by multiplying back for \(x\ne 0\): \((1/x^2)x^4=x^2\), the original numerator. The restriction \(x\ne 0\) matters because the original quotient and the simplified expression are both undefined at zero.
Worked Example: Zero Over Zero Can Lead to No Limit
For \(x\ne 0\), define \(F(x)=\frac{x\sin(1/x)}{x}\). At \(x=0\), define the numerator to be 0 and the denominator to be 0, so substitution into the displayed quotient’s numerator and denominator gives \(0/0\). Determine whether the quotient has a limit as \(x\to 0\).
For every \(x\ne 0\), cancellation gives \(F(x)=\sin(1/x)\). Consider the two sequences
Both sequences approach zero: their denominators grow without bound, so \(x_n\to 0\) and \(y_n\to 0\). But the corresponding function values are
If \(F(x)\) had a limit as \(x\to 0\), its values along every sequence approaching zero would approach that same limit. Along the first sequence the values are constantly 1, while along the second they are constantly \(-1\). Since those values do not approach the same number, the limit does not exist.
Check the cancellation for nonzero \(x\): \(\sin(1/x)\cdot x=x\sin(1/x)\), so multiplying the simplified quotient by the denominator recovers the original numerator. The calculation is valid near zero except at zero itself, as required for this limit analysis.
A Reliable First-Step Procedure
When a quotient limit is presented, begin by evaluating the numerator and denominator limits separately, if those limits are known. For polynomial expressions, direct substitution supplies these limits. Then check the denominator result before applying the quotient law.
For a polynomial quotient, use polynomial limits to compute the two values.
If its limit is nonzero, the quotient law applies. If its limit is zero, do not divide the substitution results.
If the numerator limit is also zero, record that direct substitution is inconclusive, not that the quotient limit is zero.
Try an algebraic rewrite that gives an equivalent expression for nearby inputs, or use another method appropriate to the functions involved.
State why the rewrite is valid near the point and verify that the simplified expression has the claimed behavior.
A useful algebraic rewrite need not preserve the function’s value at \(a\), where the original quotient may be undefined. It must preserve the value for inputs sufficiently close to \(a\), excluding \(a\) if necessary. Limits depend on those nearby values, not on assigning a value at the point itself.
A Common Mistake: Using Function Values as Limits
A tempting but incorrect argument is: “The numerator is zero at \(a\), the denominator is nonzero at \(a\), so the quotient limit is zero.” This conclusion does not follow for arbitrary functions unless the needed limits are known. For example, define \(f(a)=0\), define \(f(x)=\sin(1/(x-a))\) for \(x\ne a\), and let \(g(x)=1\) for all \(x\). Then \(f(a)/g(a)=0/1=0\), but \(\frac{f(x)}{g(x)}=\sin(1/(x-a))\) for \(x\ne a\), which has no limit as \(x\to a\).
The correction is to check the limit hypotheses, not merely the point values. In this example the numerator limit does not exist, so the quotient law cannot be applied. If \(f\) and \(g\) are continuous at \(a\) and \(g(a)\ne 0\), then their limits equal their values at \(a\), and the quotient law does apply. Polynomial quotients with a nonzero denominator value are an important case of this situation.
Why This Recognition Matters
Recognizing \(0/0\) prevents two opposite errors. It prevents a learner from claiming that the quotient limit is zero just because both substitution results are zero. It also prevents the opposite conclusion that the limit must fail to exist because the quotient is undefined at the point. As the examples show, neither conclusion is justified by the substitution result alone.
Instead, zero over zero identifies a decision point. The quotient law has not resolved the limit, so examine the expression for a way to rewrite it on a punctured neighborhood of the point. Factoring, cancellation, and other algebraic transformations can sometimes expose a simpler expression. Whether a particular transformation applies depends on the expression and must be checked. If no such rewrite is available, a different limit method may be needed.
Check Your Understanding
Answer each question by distinguishing the substitution result from what can actually be concluded about the limit.
- For a polynomial quotient, what does substitution giving \(0/0\) tell you, and what does it not tell you?
- State the condition on the denominator limit required by the quotient law.
- Why can two expressions that differ at the point \(a\) still be useful for evaluating a limit as \(x\to a\)?
- Give one possible behavior of a quotient after a zero-over-zero substitution, and explain how a valid rewrite could reveal it.
- Why are the point values \(f(a)=0\) and \(g(a)=1\) not, by themselves, enough to conclude that \(f(x)/g(x)\) approaches zero?
