Why Denominators Restrict a Domain
A formula can accept many different inputs, but not every algebraic expression is defined for every real input. In a rational formula, the key restriction is division: a quotient is defined only when its denominator is nonzero. With several variables, the excluded inputs may form a line, a curve, a point, or a surface rather than just one number.
The task is therefore to find the inputs that make the denominator zero, then describe all the remaining inputs. The numerator may affect the value of the formula, but it does not remove a denominator restriction.
Here, “natural domain” means that we use every real input permitted by the formula itself, with no additional restrictions declared from outside the formula. The denominator condition determines the domain: calculate where \(Q=0\), exclude those inputs, and retain all the others.
Why the Denominator Condition Is Exact
For real numbers, division \(a/b\) is defined when there is a number \(c\) such that \(bc=a\). If \(b\ne 0\), there is exactly one such number, namely \(a/b\). If \(b=0\), then \(bc=0\) for every \(c\). Thus, when \(a\ne 0\), no such \(c\) exists, and when \(a=0\), every \(c\) satisfies the equation, so there is no unique quotient. In either case, division by zero does not define a real number.
For a formula \(P/Q\), this reasoning applies after each input is substituted. At an input where \(Q\ne 0\), the quotient has a real value. At an input where \(Q=0\), it does not. This proves that the natural domain is exactly the set of inputs where the denominator is nonzero.
The notation \(\mathbf{x}\) stands for the ordered tuple of input variables. For example, when there are two variables, \(\mathbf{x}=(x,y)\); when there are three, \(\mathbf{x}=(x,y,z)\). The set on the right contains every real tuple except those that make \(Q\) zero.
Worked Example: A Linear Denominator in Two Variables
Worked Example: A Linear Denominator in Two Variables
Find the natural domain of \(f(x,y)=\dfrac{x+2y}{x-y+3}\).
Step 1: Identify the denominator. The denominator is \(x-y+3\). It is the only part of this rational formula that can prevent division from being defined.
Step 2: Find where it is zero. Solve the denominator equation:
Thus, every point on the line \(y=x+3\) makes the denominator zero, and no point off that line does.
Step 3: State the domain as a set.
Check: At \((0,0)\), the denominator is \(0-0+3=3\ne0\), so this input is included. At \((0,3)\), the denominator is \(0-3+3=0\), and \(3=0+3\), so this input lies on the excluded line. The set description correctly includes the first point and excludes the second.
Solving for one variable is sometimes a convenient way to describe the excluded set, as in the example above. It is not necessary to do so if the equation \(Q=0\) already gives a clear description. What matters is that the set specifies exactly which real inputs are excluded.
When the Excluded Set Is a Single Point
A denominator involving several squared variables can vanish at only one input. For example, a sum of real squares is zero only when every squared quantity is zero: each square is nonnegative, so a positive square cannot be cancelled by another square.
Worked Example: A Quadratic Denominator in Two Variables
Find the natural domain of \(g(x,y)=\dfrac{1}{x^2+y^2}\).
Step 1: Set the denominator equal to zero.
Since \(x^2\ge 0\) and \(y^2\ge 0\), their sum can be zero only if both squares are zero. Therefore \(x=0\) and \(y=0\). Conversely, substituting \(x=0\) and \(y=0\) gives \(0^2+0^2=0\), so the origin does make the denominator zero.
Step 2: Exclude that point.
Check: At \((1,0)\), the denominator is \(1^2+0^2=1\), so \(g(1,0)=1/1=1\) is defined. At \((0,0)\), it is \(0^2+0^2=0\), so the formula would require division by zero and is not defined there.
This example also shows why it is useful to solve the denominator equation rather than guess from the appearance of the formula. The denominator is zero at one point, not along either coordinate axis: for instance, at \((0,1)\) its value is \(0^2+1^2=1\), not zero.
Factoring the Denominator to Find Every Exclusion
If the denominator factors, its zeros can often be found by solving the simpler equations given by the factors. For real numbers, a product is zero exactly when at least one factor is zero. Therefore, if \(Q=AB\), then \(Q=0\) when \(A=0\) or \(B=0\). Both sets of solutions must be excluded.
Worked Example: A Factored Denominator in Three Variables
Find the natural domain of \(h(x,y,z)=\dfrac{z+1}{(z-xy)(x+y-2)}\).
Step 1: Identify and factor the denominator. It is already written as two factors, \(z-xy\) and \(x+y-2\). The product is zero if either factor is zero.
Step 2: Solve both factor equations.
The first equation describes the inputs on the surface \(z=xy\). The second describes the inputs on the plane \(y=2-x\), with \(z\) unrestricted by that equation. Points on either set make at least one denominator factor zero.
Step 3: State the domain by excluding both sets.
Check the two exclusions: At \((1,2,2)\), \(z-xy=2-(1)(2)=0\), so the denominator is zero and the input is excluded. At \((1,1,0)\), \(x+y-2=1+1-2=0\), so that input is also excluded. For an included input, take \((0,0,1)\). The factors are \(1-(0)(0)=1\) and \(0+0-2=-2\); their product is \(-2\ne0\). The formula is defined there, and \(h(0,0,1)=(1)/(-2)=-\tfrac12\).
The word “and” in the domain description is important: an included input must avoid both zero conditions. Equivalently, it must satisfy \(z\ne xy\) and \(x+y\ne2\). A point that satisfies just one of these nonzero conditions but fails the other is still excluded because one factor of the denominator remains zero.
Cancellation Does Not Restore an Excluded Input
A frequent source of error is simplifying a fraction and then using the simplified expression to enlarge the domain. Simplification can make a formula look defined at an input where its original denominator is zero. But the original formula still requires division by zero at that input. The natural domain belongs to the formula as given; it is found from its original denominator.
Worked Example: A Cancellable Factor
Find the natural domain of \(p(x,y)=\dfrac{x^2-y^2}{x-y}\), and compare the formula with its simplified form.
Step 1: Find the zeros of the original denominator. The denominator is \(x-y\), so it is zero exactly when \(x=y\). Hence the original formula has domain
Step 2: Factor the numerator.
For inputs with \(x\ne y\), cancellation gives \(p(x,y)=x+y\). The condition \(x\ne y\) is required for this cancellation because it ensures that the cancelled factor is nonzero.
Step 3: Check an excluded input. At \((2,2)\), the original numerator is \(2^2-2^2=4-4=0\), and the original denominator is \(2-2=0\). The original expression would be \(0/0\), so \(p\) is not defined there. The simplified expression \(x+y\) gives \(2+2=4\) at that point, but that value belongs to the simplified expression, not to the domain of the original formula.
Check an included input: At \((3,1)\), the original numerator is \(3^2-1^2=9-1=8\), and the denominator is \(3-1=2\), so \(p(3,1)=8/2=4\). The simplified expression gives \(3+1=4\), agreeing at this input, which satisfies the required condition \(x\ne y\).
The simplification is valid on the original domain, but it does not change that domain. More generally, cancellation can show that two formulas have the same values wherever the original quotient is defined. It does not, by itself, authorize adding inputs where the original denominator was zero.
A Reliable Method
For a rational formula in several variables, use the following sequence. It keeps the denominator condition explicit and helps prevent missed exclusions.
Do this before simplifying or cancelling any factors.
Find all real input tuples for which the denominator equals zero. Factor it when that makes the equation easier to solve.
If the denominator has several factors, exclude the zeros of every factor, not just one of them.
State the real input space and the exact nonzero condition, or describe the excluded set and take its complement.
Substitute an included input and each type of excluded input into the original denominator to confirm the set description.
Common Mistakes and Why They Fail
Mistake 1: Excluding only one factor's zeros. For a denominator such as \((x-y)(x+y)\), checking only \(x-y=0\) misses the inputs where \(x+y=0\). The product is zero in either case. The correction is to solve both equations and exclude the union of their solution sets.
Mistake 2: Treating a zero numerator as permission to divide by zero. If both numerator and denominator are zero, the quotient is still undefined. For example, \(0/0\) is not a real number with a uniquely determined value. The correction is to test the denominator condition independently of the numerator.
Mistake 3: Forgetting that inputs are ordered tuples. A condition such as \(x=y\) describes a set of points \((x,y)\), not a single real number. The domain must be stated using the relevant input variables and their real values.
Mistake 4: Cancelling before recording the restriction. In the cancellable-factor example, replacing the quotient by \(x+y\) and then claiming the domain is all of \(\mathbb{R}^2\) adds points that the original formula does not define. The correction is to find the domain from the original denominator first, and retain that restriction after simplification.
Why These Restrictions Matter
A domain tells us which inputs a function can actually accept. In later calculations, an excluded point or curve cannot be treated as an ordinary input merely because nearby inputs are allowed. For example, a formula with denominator \(x-y\) is unavailable on the entire line \(x=y\), not just at one selected point. Recognizing the shape of the excluded set is part of understanding the function, not an optional detail of notation.
Always connect the set description back to the denominator: every included tuple must make it nonzero, and every excluded tuple must make it zero. That two-way check is a compact test of whether the natural domain has been found exactly.
Check Your Understanding
For each question, identify the denominator zeros before stating the domain.
- Find the natural domain of \(F(x,y)=\dfrac{3x-y}{x+2y}\).
- For \(G(x,y)=\dfrac{1}{x^2+4y^2}\), which real input tuples make the denominator zero?
- Find the natural domain of \(H(x,y,z)=\dfrac{x+z}{(x-y)(z+1)}\). Which equations describe the excluded sets?
- Why does simplifying \(\dfrac{x^2-9}{x-3}\) not make the original formula defined at \(x=3\)?
- A rational formula has denominator \(5\). What is its natural domain in two real variables, assuming its numerator is defined for every real pair?