Print

Print


On Sat, 30 Dec 2017 22:38:57 +0100, taliesin the storyteller <[log in to unmask]> wrote:

>First attempt for sparwan yielded "ksavra". /vr/ because "skarva" is too
>close to Norw. "skarv", 

Why is that a problem?  Surely "false friends" add verisimilitude.

>and then /ks/ because while /sk/ is allowed in
>the onset, it is very limited in use and borrowing it directly felt
>wrong. 

Is there no /ps/ (or /sp/) onset?  How about /ts/?

Just as languages can borrow new phonemes through extended contact, they can borrow new phonotactic possibilities.  I don't know how well it fits Taruven, but the sort of scenario that suggests itself to me as possible is that, after establishment of a pattern that Germanic /sk-/ be borrowed as /ks-/, the rule is analogically generalised to borrowing /sp-/ as /ps-/, even if /ps-/ wasn't licit before.

>A la Portuguese:
>schola > escola
>
>A la Turkish:
>Smyrna > Ismir

The same strategy, these two.  You could use whatever epenthetic vowel Taruven favours in initial position.

>A la French:
>schola > école

This should I think be off the list, 'cause it's two steps -- there was an intermediate stage with /esk/, just like your Portuguese.

>There are no doubt a lot more, though a quick glance at wikipedia didn't
>turn up a quotable example of inserting a vowel between the difficult
>consonants, like so:
>
>starran > satarran

IE didn't tend to do this -- *s+stop clusters seem to have originally behaved in some ways as if unitary
  https://www.academia.edu/19044181/SPecial_STatus_presigmatised_stops
and one doesn't tend to split a unitary consonant by a vowel.  

But Persian e.g. does it!  E.g. PIE *h2ster- >> P _setâre_ 'star', Old Iranian *spaita- >> P _sefid_ 'white'.

>At this point I think I need more examples. What do your conlangs do? Do
>you have a master plan for how to transmogrify various words from
>various languages into your conlangs or do you handle each case separately?

I like the master-plan approach for any substantial donor language.  But I've never had a conlang where the number of such schemata I've needed has gotten as large as "various": it's been one or two, in the case of two maybe with one of the donors being mediated by the other.  (None of my conlangs has a setting which makes possible that they could be in contact with languages *here*.)

Alex