it's a command line parameter in espeak... IIRC (you'd have to check, though, by entering "espeak --help" in the terminal) it's -a and a numeric value between 1 and 20, with 10 being the default. you enter this in the last input field that's empty by default.

anyway, this is only half of the solution:
i found that i had to go to settings and change the volume of my ringtone so it's a little lower. the problem is that usually, you'll want to have both the ringtone and espeak to be roughly at the same level - the level you're comfortable with. doesn't work in practice. Either the ringtone isn't loud enough or it makes the spoken voice hard to understand.

Symbian has a solution for this:
the volume of the ringtone goes down just a bit during voice output. then it returns to normal until the caller name is spoken again.

i don't know if this would be possible in maemo - if one application could change the volume of the audio output of another application. because this is what espeakcaller would have to do: lower the volume of the ringtone before speech synthesis starts, then set it back to normal.