From my point of view teh problem is that Humans (which is most of us - one or two I am not sure of

) when a strange voice is heard, especially unexpectedly, we do not properly heard the first few words
I learned this when I saw a video about answering the phone - the 'trainer' said that the best way to ring someone was to start with something like
"Good Morning (pause) this is xxx from yyy "
i.e. 4 words and a pause before actually delivering any information you want the recipient to know
and that is a phone call where the other person has picked up the call so is expecting a conversation
so - a cyclist bimbling along suddenly hearing "coming up on your right" is lucky if they tune in by the last word
and the information they have is basically
a) a human voice has been heard
b) it came from behind
c) it might possibly have end in "right" - I have actually had people say "on your right" and people say "move right" and people say " don;t move right" - doesn;t really work
soooo
I have a fraction of a second to react if at all - to a person - in an unknown place behind me - and avoid a collision
it could be a pedestrian that I didn;t notice
or a person in a car
or a cyclist
dunno
but a bell - now that has an expectation of it being a cyclist - hence I can make several assumptions - like they are probably going to pass me - and they are aware of how a cyclist works and will not be a danger
and - hopefully, but not today - they will pass at a reasonable distance
which is why I have a bell - people relate a bell to an approaching bike