R: backwards conlanging

As always, I suggest the visit to one of M. Rosenfelder's pages. Set your
browser on:

http://www.zompist.com/virtuver.html

Click on the 'Proto Eastern' link and tell me what do you think!

Luca

> Query:
>
> I've read several conlang sites that suggest starting with an
> ancestor-language and then deriving conlangs from the ancestor-conlang.
>
> This is a good idea.  The problem (for me) is that I started developing
> Chevraqis before I'd read about this (or it had penetrated, anyway).
> Since then I've been struggling with the ancestor-conlang trying to get
> the thing to come out so I don't shriek in disgust whenever I look at it.
>
> Can you work backward from a single conlang to develop an ancestor-lang?
> I figure I could do that with what I have, reverse-engineer the ancestor,
> adn then resume "normal" conlang production by creating stuff in the
> ancestor and evolving it to the current language.
>
> Crowley's _Historical Linguistics_ explains about reconstruction of
> ancestor tongues, but this requires taking actual languages that really
> did probably evolve from a common ancestor and working backward from
> several of them, if I understood that chapter remotely correctly.  I
> don't want to just randomly create another language and try to work
> backwards, because in that case they *won't* have a common ancestor like
> in "real life."
>
> Does anyone have suggestions for a related-but-faux conlang technique for
> same?
>
> YHL, who is otherwise ready to scrap the entire mess and start from
> scratch, darnit
>