0

Can we find the domain of a composite function $f \circ g$ by simplifying $f \circ g$ (using identities or any rewriting rule) and then taking the intersection of the natural domain of it with $g$? Note that some simplifications change the domain of functions. Example: $$f(x)= x^2,\qquad g(x)=\sqrt x,$$ $$\operatorname{Domain}(f)=\mathbb R,\;\operatorname{Domain}(g)=\mathbb R^{\geq0},$$ $$\Rightarrow(f \circ g)(x)=(\sqrt x)^2=x,$$which has a natural domain of $\mathbb R$ after simplification. Therefore, by our method, the domain of the composite is $$\mathbb R \cap \mathbb R^{\geq0}= \mathbb R^{\geq0}$$ which can be verified using the formal definition of domain of composite functions, namely: $$\operatorname{Domain}(f\circ g)=\{x\in \operatorname{Domain}(g)\vert \,g(x)\in \operatorname{Domain}(f)\}$$ The method is usually faster than using the formal definition, and it has worked in every case I tried. But I'm not sure if it always works. Please provide a formal proof.

J. W. Tanner
  • 60,406
M.Mahdi
  • 63
  • Guys please consider the fact that we don't need the domain of the outer function f. I wrote it, so that the problem's information could be complete with well defined functions. In the method, We take the intersection of the natural domain of the composite function with the domain of the inner function. No need for the domain of the outer function 'f'. – M.Mahdi May 22 '20 at 01:55

1 Answers1

0

As you said, it is true that it works in most of the cases but not all. Actually, taking the intersection of domain of both the functions is correct but not sufficient. You should add one more point in that.

$\bullet$ Find those inputs, $x$, in the domain of $g$ for which $g(x)$ is in the domain of $f$. That is, exclude those inputs, $x$, from the domain of gf or which $g(x)$ is not in the domain of $f$. The resulting set is the domain of $f∘g$.

In your example, only taking intersection of domain of both the functions worked. But it will not in case like

Let $f(x) =\sqrt{x}$ and $g(x)=x-2$.

Here, the intersection of domain of both the functions comes out to be $[0,\infty]$. But, if you take into consideration the point which I mentioned then,

$$f\circ g=\sqrt{x-2}$$

The domain for the square root is all inputs that make $x – 2$ non-negative. That is, all $x$ such that $x – 2 \ge 0$. Solving this for $x$ gives the domain as $[2, \infty]$

You can also check this link.

SarGe
  • 3,010
  • Thanks for your respond. I guess there is a misunderstanding. The method I mentioned, actually works on your sample problem. You should first form the composite function, then after simplification (if any) take the intersection of its natural domain (the biggest domain the function is defined on) with the inner function's domain. We do not need the domain of the outer function in this method (which is actually what makes me doubt about the validity of this method). Sorry if the way I expressed my question and the method, was vague. – M.Mahdi May 22 '20 at 01:48