Skip to content

Chapter 1 — Fundamental Concepts

1.4 — Fractional Expressions

A quotient of polynomials inherits the rules of a numerical fraction, but it also inherits its caution: not every cancellation is legitimate.


In §1.3 we learned to factor. Now we use that tool in its most immediate application: quotients of polynomials. Everything §1.1 established about fractions of real numbers holds here almost unchanged. There is a single essential difference, and it shapes the whole section: the denominator is no longer a fixed number but an expression that depends on xx. The condition that it not vanish, which with numbers ruled out a single case, now rules out in general a set of values of xx, and that set must be kept in view at every step.

§1. Rational expressions and their domain

Section titled “§1. Rational expressions and their domain”

A rational expression is a quotient of two polynomials,

P(x)Q(x)\frac{P(x)}{Q(x)}

where QQ is not the zero polynomial. For each real number we put in place of xx, the numerator and the denominator become numbers, and the quotient makes sense only if the resulting denominator is not zero. We call the domain of the expression the set of reals for which it is defined:

Dom⁡(P(x)Q(x))={x∈R∣Q(x)≠0}\operatorname{Dom}\left(\frac{P(x)}{Q(x)}\right) = \{x \in \mathbb{R} \mid Q(x) \neq 0\}

The formula reads: the domain is the set of real xx such that Q(x)Q(x) is different from zero. The vertical bar, as in §1.1, separates the variable from the condition it must satisfy. The same idea applies to any algebraic expression: its domain is the set of reals for which all of its operations are defined.

To find the domain, we factor the denominator. The lemma of §1.3 says that a product of reals is zero only if one of the factors is; and if a factor is zero, so is the product. Therefore the values that make QQ vanish are exactly those that make one of its factors vanish. For example, in

x2−1x2−x−6=x2−1(x−3)(x+2)\frac{x^2 - 1}{x^2 - x - 6} = \frac{x^2 - 1}{(x - 3)(x + 2)}

the denominator vanishes at x=3x = 3 and at x=−2x = -2, and at no other value. The domain is R∖{−2,3}\mathbb{R} \setminus \{-2, 3\}; the symbol ∖\setminus indicates that we remove from R\mathbb{R} the numbers listed in braces. An irreducible quadratic factor contributes no exclusions: by §1.3, a trinomial with negative discriminant vanishes at no real number. Thus the domain of 1x2+1\dfrac{1}{x^2 + 1} is all of R\mathbb{R}.


Simplifying a rational expression consists in factoring the numerator and the denominator completely and canceling the factors they share. The rule that allows it is

P(x)R(x)Q(x)R(x)=P(x)Q(x)for all x with Q(x)≠0 and R(x)≠0\frac{P(x)R(x)}{Q(x)R(x)} = \frac{P(x)}{Q(x)} \qquad \text{for all } x \text{ with } Q(x) \neq 0 \text{ and } R(x) \neq 0

and it rests on a fact about the real numbers.

Proposition. If a,b,c∈Ra, b, c \in \mathbb{R}, with b≠0b \neq 0 and c≠0c \neq 0, then abcb=ac\dfrac{ab}{cb} = \dfrac{a}{c}.

Proof. Since b≠0b \neq 0 and c≠0c \neq 0, the lemma of §1.3 gives cb≠0cb \neq 0, so the left-hand side makes sense. Moreover, (cb)(c−1b−1)=(cc−1)(bb−1)=1(cb)(c^{-1}b^{-1}) = (cc^{-1})(bb^{-1}) = 1, so the inverse of cbcb is c−1b−1c^{-1}b^{-1}. Then abcb=ab⋅c−1b−1=ac−1(bb−1)=ac−1=ac\dfrac{ab}{cb} = ab \cdot c^{-1}b^{-1} = a c^{-1} (b b^{-1}) = ac^{-1} = \dfrac{a}{c}, using the associativity and commutativity of §1.1. ■\blacksquare

For each xx in the domain of the original expression, the polynomials PP, QQ and RR take real values aa, cc and bb that satisfy the hypotheses, and the proposition gives the equality. Outside that set the proposition says nothing, and therein lies the delicate point. Let us look at an example:

x2−1x2+x−2=(x−1)(x+1)(x−1)(x+2)=x+1x+2(x≠1,x≠−2)\frac{x^2 - 1}{x^2 + x - 2} = \frac{(x - 1)(x + 1)}{(x - 1)(x + 2)} = \frac{x + 1}{x + 2} \qquad (x \neq 1, x \neq -2)

The expression on the left is not defined at x=1x = 1, because its denominator equals 00. The one on the right is, and it equals 23\frac{2}{3}. The two expressions agree on the whole domain of the first, but they are not the same expression, because their domains differ. The simplified expression does not remember where it came from, and so whoever simplifies must carry the exclusion along explicitly, as we did in parentheses.

We say that a rational expression is simplified when its numerator and its denominator share no factor of degree greater than or equal to 11.


The product, quotient, sum and difference of rational expressions follow the rules for fractions of real numbers from §1.1. If AA, BB, CC and DD are polynomials,

AB⋅CD=ACBDAB÷CD=AB⋅DCAB±CD=AD±BCBD\frac{A}{B} \cdot \frac{C}{D} = \frac{AC}{BD} \qquad\qquad \frac{A}{B} \div \frac{C}{D} = \frac{A}{B} \cdot \frac{D}{C} \qquad\qquad \frac{A}{B} \pm \frac{C}{D} = \frac{AD \pm BC}{BD}

The justification is the same as in the previous subsection: for each admissible xx, the four polynomials take real values, and for real numbers these rules are already proved. What “admissible” means depends on the operation. In the product and the sum, the zeros of BB and DD are excluded. In the quotient we must also exclude the zeros of CC, because dividing by CD\frac{C}{D} requires that fraction not to be zero, and in the formula CC moves to the denominator.

To add or subtract, the denominator BDBD always works, but it rarely leads to a simplified result. It is better to use the least common denominator: we factor each denominator into irreducible factors (§1.3) and take the product of all the factors that appear, each with the highest exponent with which it occurs. For example, to compute

xx2−1−1x+1\frac{x}{x^2 - 1} - \frac{1}{x + 1}

we factor x2−1=(x−1)(x+1)x^2 - 1 = (x - 1)(x + 1). The least common denominator is (x−1)(x+1)(x - 1)(x + 1), and it suffices to multiply the second fraction, top and bottom, by x−1x - 1:

x(x−1)(x+1)−x−1(x−1)(x+1)=x−(x−1)(x−1)(x+1)=1x2−1(x≠±1)\frac{x}{(x - 1)(x + 1)} - \frac{x - 1}{(x - 1)(x + 1)} = \frac{x - (x - 1)}{(x - 1)(x + 1)} = \frac{1}{x^2 - 1} \qquad (x \neq \pm 1)

Multiplying top and bottom by x−1x - 1 is the simplification rule of §2 used in reverse, and so it requires x≠1x \neq 1, which was already excluded anyway.


A complex fraction is a fraction whose numerator, whose denominator, or both, themselves contain fractions. There are two ways to reduce it to a simple fraction, and both lead to the same result. The first is to work separately on the numerator and the denominator until each is a single fraction, and only then apply the quotient rule, A/BC/D=ADBC\dfrac{A/B}{C/D} = \dfrac{AD}{BC}, valid where BB, CC and DD do not vanish. The second is to multiply the main numerator and the main denominator by the least common denominator of all the inner fractions, which disappear in one stroke; again, this is the rule of §2 used in reverse.

Let us look at an example with two variables, xx and hh. By the first route,

1x+h−1xh=x−(x+h)x(x+h)h=−hx(x+h)h=−hh⋅x(x+h)=−1x(x+h)\frac{\dfrac{1}{x + h} - \dfrac{1}{x}}{h} = \frac{\dfrac{x - (x + h)}{x(x + h)}}{h} = \frac{\dfrac{-h}{x(x + h)}}{h} = \frac{-h}{h \cdot x(x + h)} = \frac{-1}{x(x + h)}

The starting expression requires x≠0x \neq 0, x+h≠0x + h \neq 0 and h≠0h \neq 0. The last step cancels the factor hh, and the proposition of §2 allows it precisely because h≠0h \neq 0. Here the warning of §2 reappears: the result −1x(x+h)\frac{-1}{x(x + h)} makes sense for h=0h = 0, but the original expression does not, so the equality holds only for h≠0h \neq 0. This quotient is not just any example. It compares how much 1x\frac{1}{x} changes when xx is shifted by hh, and Calculus studies what happens to it as hh approaches 00. That the simplification removes hh from the denominator is what makes that question possible.

Let us also note what was canceled: the factor hh, which multiplies the whole numerator and the whole denominator. A summand is never canceled.


Fractional expressions are not limited to quotients of polynomials. They may contain radicals, and then the domain has a second source of exclusions: by §1.2, a radical of even index is defined only for nonnegative radicands. For example, in xx−1\dfrac{\sqrt{x}}{x - 1} we must require x≥0x \ge 0 because of the root and x≠1x \neq 1 because of the denominator. In the interval notation of §1.1, the domain is [0,1)∪(1,∞)[0, 1) \cup (1, \infty).

In §1.2 we rationalized denominators by multiplying by the conjugate. Sometimes the opposite is convenient, rationalizing the numerator, and the tool is the same: the difference of squares from §1.3. Let us return to the quotient of the previous subsection, now with a root:

x+h−xh=(x+h−x)(x+h+x)h(x+h+x)=(x+h)−xh(x+h+x)=1x+h+x\frac{\sqrt{x + h} - \sqrt{x}}{h} = \frac{\left(\sqrt{x + h} - \sqrt{x}\right)\left(\sqrt{x + h} + \sqrt{x}\right)}{h\left(\sqrt{x + h} + \sqrt{x}\right)} = \frac{(x + h) - x}{h\left(\sqrt{x + h} + \sqrt{x}\right)} = \frac{1}{\sqrt{x + h} + \sqrt{x}}

Let us review the conditions. The starting expression requires x≥0x \ge 0, x+h≥0x + h \ge 0 and h≠0h \neq 0. In the first step we multiplied top and bottom by x+h+x\sqrt{x + h} + \sqrt{x}, and for that this number must not be zero. The conditions guarantee it: a sum of two nonnegative roots equals 00 only if both are 00, that is, if x=0x = 0 and x+h=0x + h = 0, which would give h=0h = 0, an excluded value. In the second step we used (u)2=u\left(\sqrt{u}\right)^2 = u for u≥0u \ge 0, and in the last we canceled the factor hh, which is not zero. As in the previous subsection, the result makes sense for h=0h = 0 and the original expression does not.


With the handling of algebraic and fractional expressions, the chapter now has everything it needs to take up the solution of equations, the subject of the next section.