On 2009-01-05 Roger Mills wrote:
 > An additional problem that gets a lot of
 > attention is that Prevli has inalienable
 > possession, so that in a sentence like "Mary
 > says that she will feed the baby", _baby_ MUST
 > carry a possessive suffix, usually -z, so "titi-
 > z" 'baby' can mean 'his/her baby' or (in
 > context) 'the baby'. One would assume that we
 > mean "Mary will feed (her own) baby", but what
 > if Mary is an aunt, a sitter, or just a helpful
 > friend, etc. ? (The pronoun "she" is also ambiguous--
 > Mary? someone else?) How necessary would it be
 > to be absolutely precise in cases like this? Or
 > should we just leave it to our old friend
 > "context"?

I have a strong feeling that natlangs with this
kind of ambiguity would leave it to context.

I'm reminded of the story of the Jesuit
missionaries who were worried that the word
'father' was inalienably possessed in Huron, so
that they felt that the theological semantics of
the Lords Prayer would be obfuscated. There is no
report of what the Huron converts felt, but to me
it seems like a case of reading too much semantics
into grammar.

AFMOC Kijeb has three sets of third person
pronouns called "proximal, medial, distal" or
"near, farther away, out of sight", which can be
used in a case like this, but in practice personal
and possessive endings, which would help in this
particular case since there is a reflexive
possessive, are used on their own. When subject
and object are the same person, gender and number
the verb gets only one participant marker, since
the two participants have the same rank, and the
duty of telling that there are two participants at
all falls on transitivity marking -- if that is
present at all since it relies partly on opaque
non-productive morphology (e.g. if "feed from the
breast as the baby does" is _digya_ then "feed at
the breast as the mother does" is _brigya_ from
pre-Kijeb _*pu'diya_ or something completely
different, and if the intransitive verb has an
initial /p/, /f/ or /kp)/ to begin with the
morphological transitivity marker will not be
realized at all).

I read somewhere (Roger Lass?) that language
doesn't go out of its way to resolve ambiguities
which arose through erosive language change or
which was there 'to begin with'; if a means to
resolve ambiguity should arise as a side effect of
some change speakers may seize upon it, but
otherwise language just tolerates a high degree of
ambiguity.  Neither, of course, does language
(i.e. its speakers) go out of their way to avoid
letting erosive language change make ambiguity
arise. That is language change -- in particular of
course phonetic change -- is blind, lacking both
foresight and hindsight.  Realizing this is IMHO a
big part in making fictional historical conlanging

/BP 8^)>
Benct Philip Jonsson -- melroch atte melroch dotte se
  "C'est en vain que nos Josués littéraires crient
  à la langue de s'arrêter; les langues ni le soleil
  ne s'arrêtent plus. Le jour où elles se *fixent*,
  c'est qu'elles meurent."           (Victor Hugo)