 Quote by Erland
When solving problems of this kind, one can square both sides and manipulate until no square roots remain. But when squaring, the implication goes only forward. Therefore, one must check all solutions one finds by inserting them into the original equation.
For example, squaring the equation ##\sqrt x =-1## gives ##x=1##. The first equation can therefore have no solution other than ##x=1##. But if we insert this in the first eqaution, we obtain ##\sqrt 1=-1##, which is false. Therefore, ##x=1## is not a solution either, so the first equation has no solutions.
|
I hope you took the square root function in the sense of the principal branch, for if you did not, we have that [itex]\sqrt{}[/itex] is a multivalued function, having precisely two values for every real number other than 0, and it is one-valued for 0. Taking the square root as multivalued yields [itex]-1\in \sqrt{1}[/itex], which is true, for [itex](-1)^2=1[/itex]. However, the square root in the principal branch is defined as the
positive answer to the multivalued square root, making it one-valued. In general, [itex]\sqrt[n]{}[/itex] is a multivalued function, having n values for all complex numbers and 1 for 0. For example, the real number 2 has three cube roots: [itex]\sqrt[3]{2}[/itex] which is the positive root, and two complex roots: [itex]\displaystyle -\frac{\sqrt[3]{2}}{2}+\frac{\sqrt[3]{2}\sqrt{3}}{2}i[/itex] and [itex]\displaystyle -\frac{\sqrt[3]{2}}{2}-\frac{\sqrt[3]{2}\sqrt{3}}{2}i[/itex]. We avoid this problem by defining the result of a rooting operation to satisfy these if x is real:
[itex]\sqrt[n]{x}[/itex]'s principal branch value is defined so as to be the unique
real number a satisfying [itex]a\in\sqrt[n_m]{x}[/itex] and [itex]a\geq 0[/itex], where [itex]\sqrt[n_m]{}[/itex] denotes the multivalued version of the nth root.
I just wanted to point that out to avoid confusion.