Dental sonorants ...

["Alfred W. Tüting"]

 >(John:)The shift of intervocalic s (probably always z) to r is fairly
widely
attested, e.g., in Latin, where, as I understand it, the infinitive in -re
is from an old locative in *se (or maybe it was *si).  And Norse -Vr
masculine singulars and plurals are from *-Vz.  As I understand it there a
sort of continuum from fricative to trill - trilling is exagerated
friction?  However, I don't think any of the Siouan r's are trilled, just
tapped.<<

It appears that S (Z, TH/DH), L and R are pretty close and
interchangeable (in the sense of shift) in many languages, obviously due
to the tongue tip's position:
e.g. Chinese/Japanese: L/R, Latin/Romanian: (intervocalic) L -> R
(angelus -> înger(-ul), filum -> fir(-ul) etc.), Welsh:  LL is produced
with the tongue tip in L-position and 'blowing' as if pronouncing an S.

Alfred