Wikipedia mentions the following arithmetic fallacy: $$x<0 \land y <0 \land \sqrt{x \times y}= \sqrt{x} \times \sqrt{y}, $$ since this would lead to $-1=i^2=1$.
So, the above rule is ommitted from the rules of arithmetic of complex numbers.
Now, if we instead of thinking of arithmetic operators as functions, we consider them as relations, then would it be possible to keep the above rule?!
Let's adopt the notation $$a \ * \ b \leadsto c$$ to mean: $c$ is a result of applying operator $*$ on $a, b$.
So, $a \times b \leadsto c$ means: $c$ is a result of $a \times b$.
Then we may maintain the above rule and have: $$ i^2 \leadsto c \iff [c=-1 \lor c=1]$$
Would this lead to an inconsistency in arithmetic of complex numbers?