Open lynnda-hill opened 2 years ago
So what do you want the output / analysis to be?
I would like to get:
"<feara makkárge>"
instead of:
"<feara makkár>"
After the latest changes in the clitics.lexc
file, the analysis is now:
"<feara makkár>"
"feara makkár" Pron Indef Attr <W:0.0> @OBJ> #11->11
"<ge>"
"ge" Pcle Foc/Neg-ge <W:0.0> @PCLE #12->12
"ge" Pcle Foc/Pos-ge <W:0.0> @PCLE #12->12
To avoid this split, we need to lexicalise the pronoun including the clitic. That will give us two analyses that needs to be disambiguated in the mwe-dis.cg3
file. If that is still not good, we need to remove feara makkár
from the clitics altogether.
Any preferences, @lynnda-hill and @duomdaamaendra ?
According to the lexc code, feara makkár
can take any clitic. Is this true? What is the status of ge
in feara makkárge
?
The attributive multi-word pronoun "feara makkárge" is automatically tokenized with a separate particle "ge" before any Constraint Grammar analysis. For grammar rules it is a disadvantage to have a particle between an attributive pronoun and its nominal head. It is also strange when two separate elements are analyzed as a MWE just to split the element in two.
Example sentence:
Analysis: