At 3:44:21 AM on Thursday, October 20, 2011, stlatos wrote:
> There should be no reason for any linguist to reject a
> sound change because it's optional.
Nothing except intelligence.
Optional sound changes are a methodological nightmare; at
best they are admissible only under the strictest controls.
Failure to recognize this leads to such crap as '[a]ll known
languages not currently classified as IE are actually from
one branch of IE: Indo-Iranian'.
(<
http://tech.groups.yahoo.com/group/cybalist/message/62316>)
They also (as Douglas pointed out) obviate any need to look
for real but highly non-obvious sound laws.
In fact it's not clear to me that truly optional sound
changes, as distinct from incomplete sound changes, dialect
mixtures, analogical changes, synchronic variation, etc.,
actually exist.
If I thought that it would do any good, I'd recommend you to
read Roger Lass, _Historical Linguistics and Language
Change_, especially sections 3.6 and 3.7.
Brian