Sound change and lenited forms

[Davush]

Hello,

In a proto-language, an initial unvoiced consonant usually becomes voiced when a prefix is added to it, or between vowels. This means that roots beginning with an unvoiced initial appear as voiced in nearly all circumstances as the language is strongly prefixing. Prepositions and other grammatical words also cause voicing on nouns. This means that apart from at the beginning of an utterance, unvoiced consonants aren't that common.

In daughter languages would it be unusual for the roots to 'revert' to the unvoiced form given that they usually surface as voiced?


E.g.:
čēša- (to eat) when conjugated always surfaces as jēša (mjēša, šajēša, etc.). The only time this does not occur is when it is used as a nominal due to suffixing.

Although I am happy with this in the proto-language which has very few instances of unvoiced consonants between vowels, I do not want every root in daughter languages to being with a voiced consonant but I can't think of any plausible ways to 'reverse' this effect?

Any ideas or examples would be welcome.

Thanks.

[Man in Space]

It would be unusual, yes. Unheard of, no—Kiput is thought to have gone through a devoicing of consonants between vowels at one point, but it's an incredibly rare change.

Incidentally, initial-consonant voicing across the board is actually a reasonable change.

[Frislander]

It would be unusual, yes. Unheard of, no—Kiput is thought to have gone through a devoicing of consonants between vowels at one point, but it's an incredibly rare change.

Incidentally, initial-consonant voicing across the board is actually a reasonable change.

I'll second that, or alternatively invoke analogy.

[Davush]

Thanks for the replies.

The proto-languages more or less already has initial voicing for all roots because a vowel or voiced consonant prefix causes voicing (which is how they surface 99% of the time). This means daughter-langs would not have any roots beginning with an unvoiced consonant. I am basically looking for a way to re-introduce the voiced/voiceless distinction word-initially.

[clawgrip]

I assume then that š does not voice to to ž?

If you want to introduce more voiceless consonants, one thing you can do is give the language lexical stress and then have the consonants in unstressed syllables device.

Or you could have long vowels or something else cause prenasalization of stops, then devoice regular stops and then eliminate prenasalized stops.

[Man in Space]

If you want to introduce more voiceless consonants, one thing you can do is give the language lexical stress and then have the consonants in unstressed syllables device.

The opposite is actually more realistic—unstressed consonants voicing.