I don't think it's a matter of others hating their own voices, it's having to listen to other people who can't speak clearly enough to understand over a mic.
There's also the fact that it makes a lot of the game sounds need to be turned off just to listen to all the voice work, especially if there are children screaming in the same room as you..
Plus, you can't reread directions if you didn't catch it the first time. Bad hearing then comes into play.
Then, there's people like me who can't keep a headset's mic working for more than a few months of use.
Not to mention the problems of "immersion breaking". You can no longer just have that other character's voice just in your head, the person behind the character, or the software that edits the person's voice behind the character, breaks it. Yes, I consider speech bubbles more helpful for rp game play. Mostly the reasons I see for no voice are the one's before this one, though.
You've given some valid reasons, but some of them apply to relatively few people. For instance, I rarely have trouble understanding people, unless of course they have either a lisp or a strong accent.
I don't understand why you'd need to turn off the game sounds, though. That would ruin it for me. I remember I had a class in 7th grade,(I don't remember what it was) where in one of the exercises, you had a person sitting on both sides of you, each telling a story. When the teacher called time, you had to recite both stories. I kicked ass at that.
I remember speech a lot better than text, but you do have a point there. Mostly just for "Left, right, center, down, right, up, left, squiggles" directions, though. "Turn left at the old oak tree and go until you reach the coast, then turn right and follow the coast until you reach the town"-type directions, though, shouldn't be much of a problem.
I can't hear a character's voice in my head. I hear my own, regardless of who's supposed to be speaking, unless I've heard that person speaking before. I say it helps RP, because it gives the characters recognizable voices. Say you wake up in a cave, and you hear someone speaking. In text, if you can't tell unless they tell you or you can see them. In voice, you have a chance of recognizing them. It just seems... better for me.
I can understand, however, that it just wouldn't work for some people. If the servers are divided up like I think they might be, there should be a "No Voice|PK" server, a "No Voice|RP" server, and a "No Voice|Not RP" server.