I am stumped by this behavior of some modifications to the ERG using PET and wonder if someone can shed some light on what I must be
missing.
The following lexical item is being left in the chart after lexical processing (full chart of "Some run." is available at
https://1drv.ms/u/s!Am2TXpQ1-kjshyPqQnj5QcjDBOHg)
L [183 1-2 run_n1 (1) 0 {} { : w_period_plr} {}] < blk: 0 dtrs: 58 parents: 299 >
n_pp_c-of_le
This is despite the presence of the following rule:
veto_non-noun_lfr := lexical_filtering_rule & [+CONTEXT <[SYNSEM anti_noun_synsem,TOKENS.+LIST <[+TNT.+TAGS <"NOUN">]>]>,+INPUT
<[SYNSEM [LOCAL.CAT.HEAD noun_or_ttl, PHON.ONSET con_or_voc]]>, +OUTPUT <>, +POSITION "I1 at C1"].
The +INPUT of which matches n_pp_c_of_le and the +CONTEXT of which matches the following (below) which is also left in the chart
(having several variations due to lexical rules).
I am at a loss for why the lexical filtering rule above is not removing the lexical item above. Of course I am most probably wrong
about their "matching"! Nonetheless, any insight would be sincerely appreciated. I have stared at and fiddled with it for far too
many hours.
Thank you,
Paul
P.S. the rules that discard generic LEs given natives and zero probability tokens have been discarded (otherwise the following would
have been pruned rather than, hopefully, leading to the pruning of run as a noun; in this test case)
L [250 1-2 generic_veto_noun (1) 0 {} { : w_period_plr} {}] < blk: 0 dtrs: 64 parents: 274 >
L [249 1-2 generic_veto_noun (1) 0 {} { : v_pas-prt-t_odlr w_period_plr} {}] < blk: 0 dtrs: 64 parents: >
L [248 1-2 generic_veto_noun (1) 0 {} { : v_pas-p-p_odlr w_period_plr} {}] < blk: 0 dtrs: 64 parents: >
L [247 1-2 generic_veto_noun (1) 0 {} { : v_pas-p_odlr w_period_plr} {}] < blk: 0 dtrs: 64 parents: >
L [246 1-2 generic_veto_noun (1) 0 {} { : v_pas_odlr w_period_plr} {}] < blk: 0 dtrs: 64 parents: >
L [245 1-2 generic_veto_noun (1) 0 {} { : v_psp_olr w_period_plr} {}] < blk: 0 dtrs: 64 parents: >
basic_unknown_word
[ ARGS *cons*
[ FIRST sign
[ ARGS *list*,
SYNSEM synsem
[ LOCAL local_min
[ CAT cat_min1
[ HEAD head
[ MINORS minors_basic
[ MIN #1:predsort ],
MOD *list*,
PRD bool,
AUX luk,
INV bool,
CASE case ],
VAL valence_min,
--DT token_head ],
CONT mrs_min0
[ HOOK hook
[ INDEX #2:semarg,
XARG *top*,
LTOP handle,
--SLTOP handle ] ],
AGR *top* ],
--MIN #1,
--SIND #2,
PHON phon_min,
PUNCT punctuation_min,
NONLOC non-local_min,
OPT bool,
LEX luk,
MODIFD xmod_min ],
KEY-ARG bool,
ORTH orthog
[ FIRST *top*,
REST *top*,
FROM #3:"5",
TO string,
FORM string,
LD bracket_list,
LB bracket_list,
RB bracket_list,
RD bracket_list ],
INFLECTD luk,
GENRE genre,
DIALECT dialect,
IDIOM bool,
RNAME basic_ctype ],
REST *top* ],
SYNSEM anti_noun_synsem
[ LOCAL local
[ CAT cat
[ HEAD partn
[ MINORS minors_basic
[ MIN #4:predsort ],
MOD *null*,
PRD bool,
AUX luk,
INV bool,
CASE case,
POSS -,
--BARE luk ],
VAL valence_sp
[ SPCMPS *null*,
SUBJ *null*,
SPR *null*,
COMPS *null*,
SPEC *cons*
[ FIRST anti_synsem_min,
REST *null* ] ],
MC na,
POSTHD bool,
HC-LEX luk,
HS-LEX luk,
NEGPOL luk,
--DT #5:token_head
[ +TI "<5:9>",
+LL ctype
[ -CTYPE- string ],
+TG string ] ],
CONT nom-obj
[ HOOK #6:hook
[ INDEX #7:semarg,
XARG *top*,
LTOP handle,
--SLTOP handle ],
RELS *diff-list*,
HCONS *diff-list*
[ LIST #8:*list*,
LAST #8 ],
ICONS *diff-list*
[ LIST #9:*list*,
LAST #9 ] ],
ARG-S *null*,
CONJ cnil,
AGR #7,
CTXT ctxt_min ],
--MIN #4,
--SIND #7,
PHON phon
[ ONSET unk_onset
[ --TL #10:generic_token_cons
[ FIRST token
[ +CLASS #11:alphabetic
[ +CASE non_capitalized+lower,
+INITIAL - ],
+FROM #3,
+FORM #12:"run.",
+TO "9",
+CARG "run",
+TNT tnt
[ +MAIN #13:tnt_main
[ +TAG "POS",
+PRB "0.99" ],
+TAGS *cons*
[ FIRST "NOUN",
REST *null* ],
+PRBS #14:*cons*
[ FIRST "0.00",
REST *null* ] ],
+PRED "_run/NOUN_u_unknown_rel",
+TRAIT token_trait
[ +UW +,
+IT italics,
+LB #15:bracket_null
[ LIST #16:*list*,
LAST #16 ],
+RB #17:bracket_null
[ LIST #18:*list*,
LAST #18 ],
+LD #19:bracket_null
[ LIST #20:*list*,
LAST #20 ],
+RD #21:bracket_null
[ LIST #22:*list*,
LAST #22 ],
+HD #5 ],
+ID #23:*diff-list*
[ LIST *list*,
LAST *list* ],
+TICK + ],
REST generic_token_null ] ] ],
LKEYS lexkeys
[ KEYREL relation
[ CFROM #3,
CTO #24:"9",
PRED predsort,
LBL handle,
LNK *list* ] ],
PUNCT no_punctuation_min
[ PNCTPR ppair ],
NONLOC non-local
[ REL 0-dlist,
QUE 0-dlist,
SLASH 0-dlist
[ LIST #25:*locallist*,
LAST #25 ] ],
OPT bool,
LEX +,
MODIFD xmod_min,
--BRDG luk ],
KEY-ARG bool,
ORTH orthography
[ FIRST "_veto_NOUN_",
REST *null*,
FROM #3,
CLASS #11,
TO #24,
FORM #12,
LD #19,
LB #15,
RB #17,
RD #21 ],
INFLECTD luk,
GENRE genre,
DIALECT dialect,
IDIOM bool,
RNAME basic_ctype,
C-CONT mrs_min0
[ HOOK #6 ],
ALTS alts_min,
TOKENS tokens
[ +LIST #10,
+LAST token
[ +CLASS #11,
+FROM "5",
+FORM "run.",
+TO #24,
+CARG "run",
+TNT tnt
[ +MAIN #13,
+TAGS *cons*
[ FIRST "NOUN",
REST *null* ],
+PRBS #14 ],
+PRED "_run/NOUN_u_unknown_rel",
+TRAIT token_trait
[ +UW +,
+IT italics,
+LB #15,
+RB #17,
+LD #19,
+RD #21,
+HD #5 ],
+ID #23,
+TICK + ] ] ]
I [64 () 1-2 <5:9> "" "run." { : } {NOUN 0}] < blk: 0 >
token
[ +CLASS alphabetic
[ +CASE non_capitalized+lower,
+INITIAL - ],
+FROM "5",
+FORM "run.",
+TO "9",
+CARG "run",
+TNT tnt
[ +MAIN tnt_main
[ +TAG "POS",
+PRB "0.99" ],
+TAGS *cons*
[ FIRST "NOUN",
REST *null* ],
+PRBS *cons*
[ FIRST "0.00",
REST *null* ] ],
+PRED "_run/NOUN_u_unknown_rel",
+TRAIT token_trait
[ +UW +,
+IT italics,
+LB bracket_null
[ LIST #1:*list*,
LAST #1 ],
+RB bracket_null
[ LIST #2:*list*,
LAST #2 ],
+LD bracket_null
[ LIST #3:*list*,
LAST #3 ],
+RD bracket_null
[ LIST #4:*list*,
LAST #4 ],
+HD token_head
[ +TI "<5:9>",
+LL ctype
[ -CTYPE- string ],
+TG string ] ],
+ID *diff-list*
[ LIST *list*,
LAST *list* ],
+TICK + ]
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.delph-in.net/archives/developers/attachments/20170506/c384288d/attachment-0001.html>