Can you have a negative base for a logarithm? If the answer is no, why can you not? Isn't the answer to the above just $2$?
Similarly, is $-3$ also a solution to the equation $\log _{x}\left( 9 \right)\; =\; 2$ (I know one solution is obviously 3).
Can you have a negative base for a logarithm? If the answer is no, why can you not? Isn't the answer to the above just $2$?
Similarly, is $-3$ also a solution to the equation $\log _{x}\left( 9 \right)\; =\; 2$ (I know one solution is obviously 3).
By definition for $a>0$ we have $$\log_a(x)=\frac{\log_e(x)}{\log_e(a)}$$ where $$\log_e(x)=\int_1^x\frac{dt}{t}$$
By definition $\log_{-3}(x)$ would be the $y$ such that $(-3)^y=x$. The trouble here is that $(-3)^y$ is not well-defined unless $y$ is an integer, so you won't be able to solve this for very many different $x$.
Even if you say that $(-3)^y$ is meaningful whenever $y$ is rational with odd denominator, this only gives you countably many different $x$ for which $\log_{-3}(x)$ can exist, so it won't behave very much like ordinary logarithms. (The corresponding antilogarithm wouldn't be continuous, for example).