Re: OT: /r/ - /l/ - /n/ [was Re: OT: Rant about degres Celsius]

Tom Wier wrote:
>A change of [r] to [l], or [l] to [r], is very frequent
>typologically......
(snip the very interesting remainder)

IIRC there's also Latin adjectival -al- ~ -ar- though I don't quite remember
the
condition, except that it involved presence of /l/ or /r/ in the root.(?)

ObConlang (how nice to be able to insert one!):  Kash either dissimilates
phrasal -r(V)# #rV-  (ro '2'  rongo 'hundred' > lorongo '200', or else
deletes the first /r/ (ri 'loc.' ratu 'street' > i ratu 'in the street').
Also historically: *deraw 'sun, day' > **reraw > lero

Also in sandhi, n-r > ndr, any nasal-l > nd etc.  and r-r, l-r, r-l irreg.
> -tr-
My favorite Indonesian language, Dawan ~Atoni of Timor, has earlier **r, l,
n all > n.  A major merger, since **r comes from at least 3 PAN phonemes,
**n from 2.  Many other Indo. languages have trouble with r..r, r..l, n..l
etc. sequences.

Shameless plug:  Tom, you may have been offlist when I announced the Kash
website-- http://cinduworld.tripod.com/contents.htm  (It already needs
revising, and I'm currently struggling with the Syntax section, not my
strong suit.....)