Battle Of The Personal Anecdotes never settles any internet arguments ever.
Leaving aside reliability concerns of anecdote in associating cause and effect, an anecdote could can suffice to say that alcohol does induce violence by means other than disinhibition (since if does so even once, this is true) but it could suffice (even, again, leaving aside reliability concerns) to say that it does not, because it not doing so in one particular case (or even a larger number of particular cases) does not support the argument that it cannot do so. All you can get from "it didn't happen this time" is "it doesn't always happen", not "it doesn't happen at all".
I've had personal experience with thinking X caused Y and only realizing years later that there was some unaccounted for variable that I had been unaware of at the time.
Yes, that is a source of reliability errors in associating cause with effect in anecdote, which is why I was discussing what was possible even in the semi-ideal case that such reliability errors were not a concern. Obviously, realistically, neither case can be made by anecdote, but anecdote can't even contradict one side of that discussion, because something not occurring in one case isn't inconsistent with it occurring sometimes, while something occurring even once is inconsistent with it never occurring.
Alcohol is an absolutely brain-warping drug and can create behaviors and new personalities which are light-years from how somebody was originally. I know this from personal experience seeing two close friends pass into alcoholic territory and become completely different people.