ja: prefix in context or not?
goodmami opened this issue · comments
ACE warns of duplicated type names, and often the resolution is obvious, but see the following:
prepositions.mtr:
;;;
;;; This is to prevent prepositions from occurring before words like `here'.
;;; (23-sep-10; ph)
ni_p_rel-loc_nonsp_rel := preposition_mtr &
[ CONTEXT.RELS < [ PRED place_n_rel, ARG0 #x ] >,
INPUT.RELS < [ PRED "ja:_ni_p_rel", ARG2 #x ] >,
OUTPUT.RELS < [ PRED loc_nonsp_rel ] > ].
lex-exp.mtr
;;; 彼女 は どこ に 行っ た の >> Where did she go?
ni_p_rel-loc_nonsp_rel := arg12_v_mtr &
[ CONTEXT.RELS < [ PRED "ja:place_rel", ARG2 #x ] >,
JA.RELS < [ PRED "ja:_ni_p_rel", ARG0 #x ] >,
EN.RELS < [ PRED loc_nonsp_rel, ARG0 #x ] > ].
Note how the second has ja:
prefixed to the context predicate but the first one doesn't? Which will match?
Of course, this is not the only difference between the two.
In theory both could, ja:place_rel is rewritten to place_rel at the end
of lex-exp.mtr, before prepositions.mtr.
But place_n_rel does not have an ARG 2, so I think the second rule is badly
written and should not match.
I will try to find time to test it properly later.
…On Sat, Nov 25, 2017 at 5:26 AM, Michael Wayne Goodman < ***@***.***> wrote:
ACE warns of duplicated type names, and often the resolution is obvious,
but see the following:
prepositions.mtr:
;;;;;; This is to prevent prepositions from occurring before words like `here'.;;; (23-sep-10; ph)
ni_p_rel-loc_nonsp_rel := preposition_mtr &
[ CONTEXT.RELS < [ PRED place_n_rel, ARG0 #x ] >,
INPUT.RELS < [ PRED "ja:_ni_p_rel", ARG2 #x ] >,
OUTPUT.RELS < [ PRED loc_nonsp_rel ] > ].
lex-exp.mtr
;;; 彼女 は どこ に 行っ た の >> Where did she go?
ni_p_rel-loc_nonsp_rel := arg12_v_mtr &
[ CONTEXT.RELS < [ PRED "ja:place_rel", ARG2 #x ] >,
JA.RELS < [ PRED "ja:_ni_p_rel", ARG0 #x ] >,
EN.RELS < [ PRED loc_nonsp_rel, ARG0 #x ] > ].
Note how the second has ja: prefixed to the context predicate but the
first one doesn't? Which will match?
Of course, this is not the only difference between the two.
—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub
<#2>, or mute the thread
<https://github.com/notifications/unsubscribe-auth/ABD8xn6rnJ_Cs9CJBiKneEHw3mdLMib8ks5s5zR3gaJpZM4QqMWW>
.
--
Francis Bond <http://www3.ntu.edu.sg/home/fcbond/>
Division of Linguistics and Multilingual Studies
Nanyang Technological University