Algebraic Simplification Before Taking the Limit

Why Simplify Before Taking the Limit?

A difference quotient describes the slope of a secant line through nearby points on a graph. To find a tangent slope from first principles, we study what happens to that quotient as the second point approaches the first.

For a function \(f\) defined near \(a\), the quotient with increment \(h\) is

The condition \(h\ne 0\) matters: when \(h=0\), the two points coincide and the quotient has a zero denominator. The derivative at \(a\), when it exists, is the limit of this quotient as \(h\) approaches zero through nonzero values.

Substituting \(h=0\) into the quotient is therefore not a way to calculate the limit. Instead, algebra may reveal that the quotient agrees with a simpler expression whenever \(h\ne0\). That simpler expression can have a limit as \(h\to0\).

If \(f\) is continuous at \(a\), then \(f(a+h)\to f(a)\), so the numerator \(f(a+h)-f(a)\) approaches zero as well. But continuity is necessary for that conclusion: without it, the numerator need not approach zero. In any case, simplifying the quotient requires care about where each expression is defined.

Definition and the Scope of the Algebra

For a two-sided limit, values of \(h\) must be available on both sides of zero and sufficiently close to zero. In the usual setting where \(f\) is defined on an open interval around \(a\), every sufficiently small nonzero \(h\) is allowed. A formula for \(f\) may, however, impose additional restrictions. For instance, if \(f(x)\) has a denominator \(x-2\), then \(x=2\) is excluded from its domain.

The key algebraic principle is modest but important: expressions that agree for every allowed \(h\) in some deleted interval around zero have the same limit there, if either limit exists. A deleted interval means \(0<|h|<\delta\) for some \(\delta>0\). The value at \(h=0\) is not part of the comparison.

Factoring to Expose the Increment

For polynomial expressions, expanding \(f(a+h)\) often produces terms with a common factor \(h\) in the numerator. Factoring that \(h\) is useful because the denominator already contains \(h\). The cancellation is valid at every nonzero increment, precisely the increments used in the limit.

Combining Fractions Before Canceling

When function values are fractions, subtracting them first may expose a factor in the numerator. The denominators must be combined using a common denominator; terms cannot be canceled across addition or subtraction. Any restrictions in the original fractions remain in force even after simplification.

Canceling a Factor Without Filling a Hole

Cancellation can make two formulas agree on their common domain even when one of the original formulas is undefined at a particular point. The simplified formula does not automatically change the original function's domain. That distinction matters when forming a difference quotient: each input \(a+h\) must still be in the domain of the original function.

Why Simplifying Does Not Change the Limit

Here is the reason the cancellations above are legitimate for finding limits. Suppose \(Q(h)\) is the original quotient and \(S(h)\) is a simpler expression, and suppose they agree whenever \(0<|h|<\delta\). If \(S(h)\to L\), then \(Q(h)\to L\) as well.

This argument uses agreement near zero, not agreement at zero. In fact, the original quotient is normally undefined at \(h=0\). A cancellation can produce a formula that has a value there, but that value is useful only as part of evaluating the limit; it is not a claim that the original quotient was defined at zero.

A Common Mistake: Canceling Across a Sum

Cancellation applies to common factors, not to separate terms within a sum. For example, the expression \(\frac{h+4}{h}\) does not simplify by canceling the \(h\) in the numerator against the denominator. The numerator is a sum, not a product with \(h\) as a factor. For \(h\ne0\), the correct rewriting is

That expression does not have a finite limit as \(h\to0\). By contrast, the numerator \(h^2+4h\) can be factored as \(h(h+4)\), so \(\frac{h^2+4h}{h}=h+4\) for \(h\ne0\). The factorization is what makes cancellation valid.

Another error is to cancel a factor and then treat the canceled point as if it were in the original domain. In the third example, \(\frac{x^2-4}{x-2}=x+2\) is true only when \(x\ne2\). Substituting \(x=2\) into \(x+2\) gives \(4\), but the original function is undefined there. Simplification may describe the function near an excluded point without filling that point in.

A Reliable Simplification Process

Before taking a limit of a difference quotient, use this sequence of checks:

1. Write the quotient with its increment \(h\), and state \(h\ne0\).
2. Record any additional restrictions that make \(f(a+h)\) undefined.
3. Substitute into \(f(a+h)-f(a)\) and simplify the numerator, combining fractions or factoring where appropriate.
4. Cancel only common factors, and state the restrictions under which the cancellation is valid.
5. Take the limit of the simplified expression, ensuring the allowed values of \(h\) approach zero from both sides for a two-sided limit.
6. Check that the simplified expression agrees with the original quotient for all sufficiently small allowed nonzero \(h\).

The main purpose of these steps is not to make the algebra look shorter. It is to expose the behavior of the quotient near zero without losing track of its original meaning or domain.