I am confused but I'd think the answer is no because that being the case I could, for instance, in an induction proof suppose $p(k)$ for any (an arbitrary) $k$ (let's say a natural number) and then, if a found another natural number like $k+1$ I could conclude, in particular for $k+1$ what I assumed for $k$. Because after all that's a property of an arbitrary object. I mean: If I know that "for all object with a certain property something happens" whenever I found "a particular object the certain property" I can, then, claim that "the something happens", right? Yet I've given proofs arguing so because I chose $k$ to be ARBITRARY and apparently, along with good reason, these proofs are wrong given that they were based in such a property of arbitrary objects.
If they don't mean the same, when should I use each? and What does it mean "arbitrary but fixed" in the context of a proof by induction?