Question: A not uncommon calculus mistake is to believe that the product rule for derivatives says that $(fg)'=f'g'$. If $f(x)=e^{x^2}$, determine, with proof, whether there exists an open interval $(a,b)$ and a nonzero differentiable function $g$ defined on $(a,b)$ such that this wrong product rule is true for $x$ in $(a,b)$.

Is my solution correct? Is there anything I need to improve?

My solution:

We're trying to find a function $g$ such that $(fg)'=f'g'.$ Using the product rule, we get $$f'g + fg' = f'g'.$$ Plugging in both $f(x)=e^{x^2}$ and $f'(x)=2xe^{x^2},$ we get $$2xe^{x^2}g + e^{x^2}g' = 2xe^{x^2}g'.$$ Simplifying and canceling $e^{x^2},$ we get $$\frac{g'}{g} = \frac{2x-1}{2x} = 1 + \frac{1}{2x-1}.$$ Taking the integral of both sides, $$\int \frac{dg}g = \int 1+ \frac{1}{2x-1} \, dx.$$ $$\log|g| = \frac{\log|2x-1|}{2} + x + C.$$ $$|g| = e^{\frac{\log|2x-1|}{2}}e^xe^C.$$ Letting $e^C=K,$ we get $$\boxed{g = Ke^x\sqrt{|2x-1|}},$$ Where K is a constant greater than 0. Therefore, on any interval $(a,b)$ that does not contain the value of $\frac{1}{2},$ there exists nonzero differentiable function $g$ defined on $(a,b)$ such that $(fg)'=f'g'$ is true for $x$ in $(a,b)$.

  • 2
    $\begingroup$ Looks correct to me. $\endgroup$ Commented Nov 13, 2020 at 1:19
  • 2
    $\begingroup$ Looks pretty good. The one thing I think I would improve is you've bound $C$ twice: first as the arbitrary constant of indefinite integration, and second as the result of $e^C$, since $C$, being an arbitrary constant, means $e^C$ is also an arbitrary constant. You might want to call $e^C$ as something other than $C$ again, maybe $C'$? $\endgroup$
    – Andrew L
    Commented Nov 13, 2020 at 1:19
  • 1
    $\begingroup$ Maybe I'll use lowercase c and uppercase C. $\endgroup$
    – CSS Jowoo
    Commented Nov 13, 2020 at 1:20
  • 2
    $\begingroup$ Best to keep your constants uppercase. Perhaps $C$ and $D$ or $C_1$ and $C_2$. $\endgroup$ Commented Nov 13, 2020 at 1:55

1 Answer 1


I think you have gone far beyond the requirements of the question, which asked merely to prove that there exists an open interval and a non-zero differentiable function with the stated properties.

To prove that such an interval and function exist, you can simply choose an open interval -- for example, $(a,b) = (1,2)$ -- and one function on that interval, for example $g(x) = e^x \sqrt{2x - 1}.$ Then show that $f(x)$ and $g(x)$ (and therefore also $(fg)(x)$) are defined on your chosen interval and calculate $f'$, $g'$, and $(fg)'$:

\begin{align} f'(x) &= \frac{\mathrm d}{\mathrm dx} e^{x^2} = 2x e^{x^2}, \\ g'(x) &= \frac{\mathrm d}{\mathrm dx}\left(e^x \sqrt{2x - 1}\right) = \frac{2x e^x}{\sqrt{2x - 1}}, \\ (fg)'(x) &= \frac{\mathrm d}{\mathrm dx} \left(e^{x^2}\cdot e^x \sqrt{2x - 1}\right) \\ &= \frac{4 x^2 e^{x^2 + x}}{\sqrt{2x - 1}} \\ &= 2x e^{x^2} \cdot \frac{2x e^x}{\sqrt{2x - 1}} \\ &= f'(x)\cdot g'(x). \end{align}

Note that it was not necessary to take the absolute value of $2x-1$ in any of these formulas, since $2x - 1 > 0$ everywhere in the domain of $f$ and $g$ as defined in this proof.

And that is essentially the proof of existence of an interval and a function with the desired properties. You could choose the interval $\left(\frac12,\infty\right)$ rather than $(1,2)$ for the existence proof (as I would) since it is one of the two maximal intervals on which such a function $g$ is defined, and perhaps you might like to show off your clever method of identifying the entire family of functions that might be the desired function $g$ (as I likely would), but all of that is bonus material beyond the plain meaning of the question.


You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .