Nur-ellen mutations (was: Re: Animacy in active languages)

Jörg Rhiemeier skrev:

> well-known-to-be-inanimate noun (such as **_tring_ as an agentive of
> _dring_ "hammer" - if there really was such an animate noun, its
> objective would indeed be _dring_) in order to force it into an agentive
> slot.  Some of these "poetic agentives" might be irregular or even
> suppletive.
This just made me come up with a question on Nur-ellen mutations.

Are there mutations for every phoneme that a noun may begin
with? And if so, does it sometimes produce a result that is
the same as a non-mutated noun begins with?

Examples:

i) _r_ mutates to _r_, i.e. not at all.

ii) Agentive _caur_ -> objective _gaur_ which in turn
    is also an agentive form _gaur_ of a completely
    different word, which in turn mutates to - i don't
    know - _ghaur_?

Are there such cases?

Or perhaps agentive forms always start with a voiceless
phoneme and objective always with a voiced? And in that
case, what happens with words beginning with vowels?

Daniel