I came across this example of how you could end up with an extraneous solution but I was wondering how it arose. We have the equation: $$x^2+x+1=0 $$ Since x=0 does not satisfy the equation, you can divide by x on both sides which yields: $$x+1+\frac{1}{x}=0$$ which is equivalent to our first equation. From our first eqution we can conclude that: $$-x^2=x+1$$ We now substitute that into the second equation to get:$$x^2=\frac{1}{x}$$ which results in $$x^3=1$$ which is equivalent to our previous equation since x cannot be 0. However, one solution from our last equation is x=1, which is not a solution to our original equation. I have a vague idea that it may have to do with the fact that you get a cubic equation and you began with a quadratic, and that steps imply the following and not vice versa, but can you provide a very detailed answer as to why it arises? Can you please provide more examples?

  • 2
    $\begingroup$ Take a look this: mathforum.org/library/drmath/view/69477.html $\endgroup$
    – Ryan Soh
    Aug 16 '20 at 9:10
  • $\begingroup$ You should make than an answer if you can. $\endgroup$
    – Toby Mak
    Aug 16 '20 at 9:11
  • 1
    $\begingroup$ It's a duplicate but i am not able to find the older post $\endgroup$
    – Arjun
    Aug 16 '20 at 9:14
  • 8
    $\begingroup$ If $A(x)=B(x)$ has the same solutions as $A(x)=C(x)$, it does not follow that $B(x)=C(x)$ also has exactly those solutions. E.g., $x=2$ has the same solution as $x=2$, but $2=2$ is satisfied by all values of $x$. $\endgroup$ Aug 16 '20 at 9:17

If we call $A(x)=x^2+x+1$ and $B(x)=x+1+\frac1x$, we can schematize your passages as such: $$A(x)=0\Leftrightarrow \begin{cases}x\ne 0\\ B(x)=0\end{cases}\Leftrightarrow \begin{cases}x\ne 0\\ A(x)=0 \\B(x)=0\end{cases}\stackrel{!!!}\Rightarrow \begin{cases}x\ne 0\\ B(x)-A(x)=0\end{cases}$$

Whereas to preserve equivalence you should have kept $A(x)=0$ in $\begin{cases}x\ne0\\ B(x)-A(x)=0\\ A(x)=0\end{cases}$

  • $\begingroup$ When you say x is not equal to zero, is it just saying if x is not 0 then blablabla... is it just a remainder? Could we avoid it since they are equal for all x in the domain of A? (of course you have to consider if it is 0 before dividing but I'm thinking whether you can omit that). $\endgroup$ Aug 16 '20 at 9:40
  • 1
    $\begingroup$ I am saying that the condition $[A(x)=0]$ is equivalent to the condition $[x\ne 0\land B(x)=0]$. What I have written is a schematization of the relevant passages of your work (explicit or implicit). How you prove (co-)implications is mostly a communicative aspect: for instance, strictly speaking I do not need someone to explain why $x^2+x+1=0$ and $\begin{cases} x\ne0\\ x+\frac1x+1=0\end{cases}$ are equivalent, because I can figure out by myself the few standard passages of interest. $\endgroup$
    – user239203
    Aug 16 '20 at 11:25
  • 1
    $\begingroup$ When I'm doing a system of inequations, I like to keep existence conditions of the terms inside the system itself as much as I can (it becomes somehow cumbersome with trigonometric functions, though). I have my reasons to believe this is a good course of action. $\endgroup$
    – user239203
    Aug 16 '20 at 11:30

This substitution ($x+1=-x^2$) expands a set of roots of the equation

because $-x^2$ also depends on $x$.

You can substitute $x+1=y$, for example.

More example, when a similar substitution gives similar problems.

Let we need to solve $$\sqrt[3]{2x+1}+\sqrt[3]{x+1}=\sqrt[3]{x-1}.$$

We obtain: $$\left(\sqrt[3]{2x+1}+\sqrt[3]{x+1}\right)^3=x-1$$ or $$2x+1+x+1+3\sqrt[3]{2x+1}\cdot\sqrt[3]{x+1}\left(\sqrt[3]{2x+1}+\sqrt[3]{x+1}\right)=x-1.$$ Now, since $$\sqrt[3]{2x+1}+\sqrt[3]{x+1}=\sqrt[3]{x-1},$$ which can get something bad, we obtain: $$3\sqrt[3]{(2x+1)(x+1)(x-1)}=-3-2x$$ or $$x(440x^2+630x+189)=0$$ and we got as one of options $x=0$.

Easy to see that $0$ is not a root of the starting equation and it happened

because we used a not correct substitution $\sqrt[3]{2x+1}+\sqrt[3]{x+1}=\sqrt[3]{x-1}.$

Now, we need to check that all roots of the equation $440x^2+630x+189=0$ are roots of the starting equation, which is not so easy.

If we want to avoid from these problems, so we need to use the following identity. $$a^3+b^3+c^3-3abc=(a+b+c)(a^2+b^2+c^2-ab-ac-bc)$$


All transformations of an equation must be reversible. With $x=0$,

$$x^2+x+1=0\leftrightarrow x+1+\frac1x=0$$ is fine.

But combining two equations in one $$\begin{cases}x+1=-\dfrac1x\\x+1=-x^2\end{cases}\leftrightarrow x^2=\frac1x$$ is not.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.