;;; -*- Mode: TDL; Coding: utf-8 -*- ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;; ;;; author | date | modification ;;; ------------------|-------------|------------------------------------------ ;;; Francis Bond (FCB)| | Emily Bender (ERB), Melanie Siegel (MS) ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;; ;;; ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;; ; File for spelling preserving lexical rules, i.e., instances ; of (subtypes of) const-ltol-rule or const-ltow-rule. ; %suffix should not appear in this file (see irules.tdl). ;;;;;;;;;;;;;;;;;;;;;;;;N-> QUANTIFIED N;;;;;;;;;;;;;;;;;;;; ; (ERB 2001-08-16) Changed the following: ; ; ARGS.FIRST.SYNSEM.LOCAL.CONT.KEY.LBL #h ; ; to ; ; ARGS.FIRST.SYNSEM.LOCAL.CONT.LBL #h ; ; So that it works properly with NumCl phrases. ; (ERB 2001-09-19) Somewhat surprisingly, this isn't restricted ; to things that are LEX +. For the moment, I'll just have it ; copy up the LEX value. Perhaps it should be restricted to ; LEX + daughters. ; (ERB 2001-09-20) Of course it can't be restricted to things ; that are LEX +, because it has to apply to NumCl phrases that ; are built up of an integer and a NumCl. But copying up the ; LEX value is still a good idea. ; (ERB 2001-09-24) Experimenting with forcing this to apply ; after all other valence requirements are satisfied. (I'm ; getting spurious ambiguity in the date expressions otherwise.) ; (MS 2002-09-05) Inserted UNSAT + on the daughter to prevent the rule to be cyclic (actually a hint from Dan F.) ;;; FCB this shouldn't be a lexical-rule (it applies after syntactic rules) quantify-n-lrule := word2word-rule & [SYNSEM [LOCAL [CAT [VAL saturated & [UNSAT -]]], LEX #lex], C-CONT [RELS , HCONS ], ARGS.FIRST.SYNSEM [LOCAL [CAT [HEAD noun_head, VAL [UNSAT +, SPR opt-1-arg & [FIRST [OPT + , LOCAL.CAT.HEAD.KEYS.KEY #rel]], SUBJ olist, COMPS olist]], CONT.HOOK [LTOP #h, INDEX #i], NUCL nucl_plus], LEX #lex]]. ; (MS 2005-01-06) Floated numeral classifiers, counting subjects. ; (MS 2005-01-10) These should copy the RELS from their ARGS, because ; measuring classifiers might have quite complicated ARGS: りんごを四キロ買った num-cl-sbj-float-lrule := num-cl-sbj-float-rule. num-cl-obj-float-lrule := num-cl-obj-float-rule. ;;;;;;;;;;;;;;;;;;;;;;;;VN->N;;;;;;;;;;;;;;;;;;;; ;; (MS 2005-05-24) These cause ambiguity in all cases of ;; N no VN wo suru ;; Try: 中国語の勉強をする vn2n-det-lrule := vn2n-rule-detarg. vn2n-nspec_arg1-lrule := vn2n-rule-nspecarg1. vn2n-nspec_arg2-lrule := vn2n-rule-nspecarg2. ;;;;;;;;;;;;;;;;;;LEXICAL RULES FOR ZPRO INSERTION;;;;; ; (ERB 2001-09-03) Now that light verbs take over the ; valence of their SPRs, I need to restrict light verbs ; from undergoing these rules. ; (MS 2003-12-22) Matrix 0.6: LKEYS.KEYREL is not accessible. ; Moving to XARG instead. ; (MS 2005-01-06) Copy up ARG-S. subj-zpro-ins-lrule := zpro-insertion-rule & [SYNSEM [LOCAL [CAT.VAL [SUBJ null, COMPS #comps, SPR #spr], CONT.HOOK.XARG zpro_ref-ind, ARG-S #args]], ARGS.FIRST [SYNSEM.LOCAL [CAT [HEAD verb_head-avm & [ LIGHT -, EMPTY -], VAL [SUBJ opt-1-arg, COMPS #comps, SPR #spr]], ARG-S #args]]]. ; (MS 2003-12-22) Matrix 0.6: LKEYS.KEYREL is not accessible. ; Moving to CONT.RELS.LIST.FIRST.ARG2. But is that right??? ; (MS 2004-02-25) No. LKEYS.KEYREL IS accessible (we are in lexrules!) ; (MS 2004-06-25) Auxiliaries should not undergo this. ; (MS 2004-11-05) why not? 食べてもらう ; (MS 2004-11-09) Ok, tried to give AUX a more complex value. Aspect ; auxiliaries should not undergo the rule, while complex auxiliaries should be ; allowed. ; (MS 2004-11-30) We need to be able to bind the second complement first: ; テーブルに置く . I have therefore changed the rule such that it can apply to a ; COMPS list that contains one or more arguments and puts the rest on the mother's list. obj-zpro-ins-lrule := zpro-insertion-rule & [SYNSEM [LOCAL [CAT.VAL [SUBJ #subj & [FIRST.OPT bool], COMPS null, SPR #spr], ARG-S [FIRST #subj]], LKEYS.KEYREL.ARG2 zpro_ref-ind], ARGS.FIRST [SYNSEM.LOCAL [CAT [HEAD verb_head-avm & [ LIGHT -, AUX aux_minus_or_complex, EMPTY -], VAL [SUBJ #subj, COMPS opt-1-arg, SPR #spr]]]]]. ; (ERB 2001-10-08) while OBJ and SUBJ zpro rules were strictly ; ordered with respect to each other, OBJ2 wasn't. I'm going ; to require it to apply first. ; (MS 2001-10-09) Now we cannot get perspective auxiliaries. We should find another solution. ; (MS 2003-12-22) Matrix 0.6: LKEYS.KEYREL is not accessible. ; Moving to CONT.RELS.LIST.FIRST.ARG2. But is that right??? ; (MS 2004-11-30) To restrict the application of this rule, it applies only if there is ; a first complement left. obj2a-zpro-ins-lrule := zpro-insertion-rule & [SYNSEM [LOCAL [CAT.VAL [SUBJ < #subj >, COMPS 1-list & < [LOCAL #obj] >, SPR #spr], ARG-S < #subj , [LOCAL #obj] >], LKEYS.KEYREL.ARG2 zpro_ref-ind], ARGS.FIRST [SYNSEM.LOCAL [CAT [HEAD verb_head-avm & [ LIGHT -, EMPTY -], VAL [SUBJ < #subj >, COMPS opt-2-args & < [OPT + ], [LOCAL #obj] >, SPR #spr]]]]]. obj2b-zpro-ins-lrule := zpro-insertion-rule & [SYNSEM [LOCAL [CAT.VAL [SUBJ < #subj >, COMPS obl-1-arg & < [LOCAL #obj] >, SPR #spr], ARG-S < #subj , [LOCAL #obj] >], LKEYS.KEYREL.ARG3 zpro_ref-ind], ARGS.FIRST [SYNSEM.LOCAL [CAT [HEAD verb_head-avm & [ LIGHT -, EMPTY -], VAL [SUBJ < #subj >, COMPS opt-2-args & < [LOCAL #obj] , [OPT + ] >, SPR #spr]]]]]. ; (CH 2005-03-18) added to allow "彼 が [zero 銀を] 金 に し た" to ; be interpreted as a v-soc-stem-lex type. obj2-soc-suru-zpro-ins-lrule := zpro-insertion-rule & [SYNSEM [LOCAL [CAT.VAL [SUBJ < #subj >, COMPS obl-1-arg & < [LOCAL #obj] >, SPR #spr], ARG-S < #subj , [LOCAL #obj] >], LKEYS.KEYREL.ARG2 zpro_ref-ind], ARGS.FIRST [SYNSEM.LOCAL [CAT [HEAD v-stem_head & [ LIGHT -, EMPTY -, COP -], VAL [SUBJ < #subj >, COMPS < [OPT + ], [LOCAL #obj, OPT -] >, SPR #spr]]]]]. ; (MS 2003-12-22) Matrix 0.6: LKEYS.KEYREL is not accessible. ; Moving to XARG instead. naadj-subj-zpro-ins-lrule := naadj-zpro-insertion-rule & [SYNSEM [LOCAL [CAT.VAL [SUBJ null, COMPS #comps, SPR #spr], ARG-S < >, CONT.HOOK.XARG zpro_ref-ind] ], ARGS.FIRST.SYNSEM.LOCAL [CAT [HEAD na-adj_head & [ADJ +, MOD null], VAL [SUBJ opt-1-arg, COMPS #comps & null, SPR #spr & null]]]]. ; (MS 2003-12-22) Matrix 0.6: LKEYS.KEYREL is not accessible. ; Moving to XARG instead. opend-subj-zpro-ins-lrule := zpro-insertion-rule & [SYNSEM [LOCAL [CAT.VAL [SUBJ null, COMPS #comps, SPR #spr], ARG-S #rest, CONT.HOOK.XARG zpro_ref-ind]], ARGS.FIRST [SYNSEM.LOCAL [CAT [HEAD v-pass-caus-end_head & [ LIGHT -, EMPTY -, MODUS operative], VAL [SUBJ opt-1-arg & < #sbj >, COMPS #comps, SPR #spr]], ARG-S [FIRST #sbj, REST #rest]]]]. ; (MS 2003-12-22) Matrix 0.6: LKEYS.KEYREL is not accessible. ; Moving to CONT.RELS.LIST.FIRST.ARG2. But is that right??? opend-obj-zpro-ins-lrule := zpro-insertion-rule & [SYNSEM [LOCAL [CAT.VAL [COMPS null, SUBJ #subj & [FIRST.OPT bool], SPR #spr], ARG-S #subj], LKEYS.KEYREL.ARG2 zpro_ref-ind], ARGS.FIRST [SYNSEM.LOCAL [CAT [HEAD v-pass-caus-end_head & [ LIGHT -, EMPTY -, MODUS operative], VAL [SUBJ #subj, COMPS opt-1-arg, SPR #spr]]]]]. ; (MS 2003-12-22) Matrix 0.6: LKEYS.KEYREL is not accessible. ; Moving to CONT.RELS.LIST.FIRST.ARG2. But is that right??? opend-obj2-zpro-ins-lrule := zpro-insertion-rule & [SYNSEM [LOCAL [CAT.VAL [COMPS 1-list & < #obj >, SUBJ #subj & [FIRST #sbj & [OPT bool]], SPR #spr], ARG-S < #sbj , #obj >], LKEYS.KEYREL.ARG2 zpro_ref-ind], ARGS.FIRST [SYNSEM.LOCAL [CAT [HEAD v-pass-caus-end_head & [ LIGHT -, EMPTY -, MODUS operative], VAL [SUBJ #subj, COMPS opt-2-args & < #obj , [OPT + ] >, SPR #spr]]]]]. ;;;;;;;;;;;;;;;;;;COPULA -> ELLIPTICAL VERB;;;;; ;cop2ell-infl-rule := ;cop2ell-rule. ;;;;;;;;;;;;;;;;;;GA-NO CONVERSION IN RELATIVE CLAUSES AND INFINITIVE+KATA;;;;; ; (MS 2003-02-14) took the restriction MOD synsem out to allow also the infinite+kata rule: ; あの人の歩き方 ; (MS 2003-08-22) added restriction obl-1-arg to ga-no-conversion and obl-arg to wo-no-conversion, ; such that spurious ambiguity in connection with VN2N rule vanishes. ga-no-conversion-rule := subj-casechange-rule & [SYNSEM.LOCAL.CAT.VAL.SUBJ obl-1-arg & [FIRST [LOCAL.CAT.HEAD.CASE no-case]], ARGS < [SYNSEM.LOCAL [CAT [HEAD verb_head & [MAIN-PRD -, COP -, H-TENSE present-or-past]], BAR -], INFLECTED +, J-NEEDS-AFFIX bool] >]. wo-no-conversion-rule := obj-casechange-rule & [SYNSEM.LOCAL [CAT.VAL.COMPS obl-arg & [FIRST [LOCAL.CAT.HEAD [PTYPE ga, CASE no-case], OPT - ]], BAR +], ARGS < [SYNSEM.LOCAL [CAT [HEAD verb_head & [MAIN-PRD -, COP -, VN -, MODUS infinitive]], BAR -], INFLECTED +, J-NEEDS-AFFIX -] >]. ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;; ;(CH 2005-04-09) ;; ;; V embedding `Alternation' Lexical Rules ;; ;; intransitive --> monotransitive ;; --> ditransitive ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;; vbar-monotransitivization-v-lrule := vbar-monotransitivization-v-rule. vbar-ditransitivization-v-lrule := vbar-ditransitivization-v-rule. vbar-monotransitivization-c-lrule := vbar-monotransitivization-c-rule. vbar-ditransitivization-c-lrule := vbar-ditransitivization-c-rule. ; vbar-monotransitivization-c2-lrule := vbar-monotransitivization-c2-rule. ; vbar-ditransitivization-c2-lrule := vbar-ditransitivization-c2-rule. ; vbar-monotransitivization-kurusuru-lrule := vbar-monotransitivization-kurusuru-rule. ; vbar-ditransitivization-kurusuru-lrule := vbar-ditransitivization-kurusuru-rule.