Noun phrase reduction

Question

A user asks:

I'm sorry if this is not the place to post my questions, but I can't login to nor create an account on support.bible.

In M, we have noun phrase reduction, which means that words ending in the 3s.POSS suffix /-na/ change their endings to /-n/ when preceding another word beginning with /t, n, l/ and /-ng/ elsewhere (M noun phrases are left-headed). For example,

numa-na                  +         sabanga     +      sibuna
house-3s.POSS                  big                       very

becomes

numang sabanga sibuna 'his very large house'

Also

buragina             +                to
morning                                first

becomes

buragin to 'wait until morning'

Because of this, I had to create three possessive suffixes /-na, -n, -ng/ in the lexicon to get words to parse correctly, calling them 3s.POSS, 3s.POSS2 and 3s.POSS3, respectively. This takes care of most of the issue, but introduces a problem, because in the target language U, they never reduce to /ng/, only /n/.

Which brings me to my first question. Is this where I create a rule to map 3p.POSS3 in M to 3p.POSS2 in U?

Now, on to my second question. In the second example above, for words without possessive suffixes on the end of them, I've entered allomorphs so that the words parse correctly (buragin being an allomorph of buragina). There are several such words I had to do this to. But FLExTrans doesn't know which word to choose in the target lexicon. What's the best way to handle this? Should I create a variant instead? A different entry altogether?

Answer 1

Question 1: yes, you want a rule to turn 3p.POSS3 into 3p.POSS2. Something like the following:

Where c_n_poss is defined as something like: n.3p_POSS, n.3p_POSS2, n.3p_POSS3, … and a_poss_slot contains all the possessive suffixes. If other possessive suffixes do something similar you could have additional when statements.

Question 2:

For this one, if it is just several words that have the ending changed, I would recommend you have two entries for each of these. Then in a noun – adjective rule, check what the adjective begins with then check what word you have and depending on the adjective output one or the other.

If it’s a lot of words that would be too many to list in a choose statement, then you would need a different approach such as artificially treating the word as if it has a suffix and putting the right suffix on it depending on the adjective.