thing (top). handle (thing). individual (thing). event (individual). sem (top) < HANDEL > = handle. ;;; ref-ind (individual). index (thing) < AGR > = agr < QUALIA > = nomqualia. agr (top) < pers > = person < num > = number < gender > = gender. gender (top) (OR male female neuter). person (top) (OR 1 2 3). number (Top) (OR sg pl). ;;; Playing with thesis stuff in mrs ... nomqualia (Top) < form > = form < telic > = verb_rel. animal (nomqualia) < form > = normal_ind. comestible (nomqualia). c_natural (comestible) . ref-ind (index). it-ind (index). there-ind (index). ; The basic semantic structure, a CONT, is just something with a handel. ; Conts are further classified into mrss (i.e., things with liszts) and ; relations (i.e., things that go on liszts). Conts are divided into psoas ; (with an distinguished event) and nom-objs (with a distinguished index). ; ; instead of CONT, we've got sem mrs (sem) < INDEX > = thing < LISZT > = diff-list. empty-sem (mrs) < LISZT : LST > = < LISZT : LAST >. psoa (mrs) < INDEX > = event. nom-obj (mrs) < INDEX > = index . common-nom-obj (nom-obj). relation (sem). ; Relations are classified according to the types of arguments they take. All ; relations have a handel. In addition, quantifier relations have a bound ; variable and a restriction, nominal relations have an instance, and event ; relations have an event. Furthermore, nominal relations and event relations ; can have addition roles, depending on their meaning. quant_rel (relation) < BV > = ref-ind < RESTR > = handle . the_rel (quant_rel). which_rel (quant_rel). ;;; More playing with thesis stuff ;;; almost certainly don't really want to do plurals and ;;; quantisation this way ... basic_nom_rel (relation) < INST > = ref-ind < QUANT > = boolean < PLMOD > = boolean. nom_rel (basic_nom_rel). tempnom_rel (basic_nom_rel). day_rel (tempnom_rel). non_day_rel (tempnom_rel). verb_prep_rel (relation). event_rel (verb_prep_rel) < EVENT > = event . temp_prec_rel (event_rel) < BEFORE > = event. no_role_rel (event_rel). ; E.g., weather verbs ;;; MRS translation of Sanfilippo style semantics ;;; do verbs differently as in Ant's version - all arguments are ;;; related by thematic role types to the event ;;; arg_rel (event_rel) < ARG > = thing . role_rel (arg_rel) < ARG > = ref-ind . verb_rel (event_rel) "rel for the verb-pred itself". ;;; roles (obviously Ant used lots lots more!) p-agt_rel (role_rel). p-pat_rel (role_rel). prep_rel (role_rel). ;;; verb semantic types verb-sem (psoa) < LISZT : LST : HD > = verb_rel < INDEX > = < LISZT : LST : HD : EVENT > < HANDEL > = < LISZT : LST : HD : HANDEL >. ;;; ignore pleonastic subjects etc for now. theta-sem (verb-sem) < LISZT : LST : TL : HD > = role_rel < LISZT : LST : TL : HD : EVENT > = < LISZT : LST : HD : EVENT > < LISZT : LST : TL : HD : HANDEL > = < LISZT : LST : HD : HANDEL >. ;;; The following are minimal adaptations of Antonio's types - ;;; it might be possible to do this more neatly strict-intrans-sem (theta-sem) "The formula which describes the semantics of a strict intransitive verb is exactly as 'verb-sem', with the exception that the second argument of the liszt encodes a thematic formula whose predicate instantiates a proto-agent role. In linear notation: [e1] [V-Pred(e1), p-agt(e1,x)]" < LISZT : LST : TL : HD > = p-agt_rel < LISZT : LST : TL : TL > = < LISZT : LAST >. trans-or-intrans-sem (theta-sem) "Intermediate type. The type 'trans-or-intrans-sem' describes the semantics of any verb which is associated with at least two formulae --- either two thematic formulae, or a thematic formula and a formula describing the semantics of a 'comp' (e.g. sentential complement) or 'xcomp' (e.g. VP complement). The first argument role is always a formula whose predicate is either a proto-agent or non thematic; the latter specification is used for extraposition and subject raising verbs where the subject argument is non thematic, e.g. 'John seems to be tired' 'It bothers Bill that Mary snores' In linear notation: [e1][V-Pred(e1), p-agt-or-no-theta(e1,x), Formula]". trans-or-intrans-no-comp-or-xcomp-sem (trans-or-intrans-sem) "Intermediate type. 'trans/intrans-no-comp/xcomp-sem' characterizes the semantics of verbs of type 'trans/intrans-sem' which have no 'comp' or 'xcomp' complements. 'Bill thought that Mary would come'. 'trans/intrans-no-comp/xcomp-sem' verbs have a thematic subject. In linear notation: [e1][V-Pred(e1), p-agt(e1,x), ??))" < LISZT : LST : TL : HD > = p-agt_rel. strict-trans-sem (trans-or-intrans-no-comp-or-xcomp-sem) "The semantics of a strict transitive inherits all specifications from the type 'trans/intrans-no-comp/xcomp-sem', and instantiates a proto-patient formula as the second and last argument role. In linear notation: [e1][V-Pred(e1), p-agt(e1,x), p-pat(e1,y)]" < LISZT : LST : TL : TL : HD > = p-pat_rel < LISZT : LST : TL : TL : HD : EVENT > = < LISZT : LST : HD : EVENT > < LISZT : LST : TL : TL : HD : HANDEL > = < LISZT : LST : HD : HANDEL > < LISZT : LST : TL : TL : TL > = < LISZT : LAST >. intrans-obl-sem (trans-or-intrans-no-comp-or-xcomp-sem) "For intransitive verbs that have a prepositional complement, e.g. 'John talked to Bill'. The second and last argument role is a prep_rel In linear notation: [e1][V-Pred(e1), p-agt(e1,x), prep(e1,y)]" < LISZT : LST : TL : TL : HD > = prep_rel < LISZT : LST : TL : TL : HD : EVENT > = < LISZT : LST : HD : EVENT > < LISZT : LST : TL : TL : HD : HANDEL > = < LISZT : LST : HD : HANDEL > < LISZT : LST : TL : TL : TL > = < LISZT : LAST >. intrans-or-trans-obl-trans-or-ditrans-sem (trans-or-intrans-sem) "Intermediate type. This type, a subset of 'trans/intrans-sem', characterizes the semantics of verbs which take three complements. In linear notation: [e1][V-Pred(e1), p-agt-or-no-theta(e1,x), ??". trans-or-obl-trans-or-ditrans-sem (intrans-or-trans-obl-trans-or-ditrans-sem) "Intermediate type. For transitive and ditransitive verbs. The argument list of the conjunctive formula which follows the subject argument role in 'intrans/trans/intrans-sem' has as its first member a thematic formula which instantiates either a proto-patient or a non thematic role; the latter specification is for object raising verbs, e.g. 'Jon believes Mary to be right'. In linear notation: [e1][V-Pred(e1), p-agt-or-no-theta(e1,x), p-pat-or-no-theta(e1,y), Formula]". obl-trans-or-ditrans-sem (trans-or-obl-trans-or-ditrans-sem) "For transitives taking an oblique complement and ditransitives, e.g 'Mary gave a book to Bill' 'Mary gave Bill a book'. The object role is thematic and it is followed by a prepositional formula. In linear notation: [e1][V-Pred(e1), p-agt(e1,x), p-pat(e1,y),prep-pat(e1,z)]" < LISZT : LST : TL : HD > = p-agt_rel < LISZT : LST : TL : HD : EVENT > = < LISZT : LST : HD : EVENT > < LISZT : LST : TL : HD : HANDEL > = < LISZT : LST : HD : HANDEL > < LISZT : LST : TL : TL : HD > = p-pat_rel < LISZT : LST : TL : TL : HD : EVENT > = < LISZT : LST : HD : EVENT > < LISZT : LST : TL : TL : HD : HANDEL > = < LISZT : LST : HD : HANDEL > < LISZT : LST : TL : TL : TL : HD > = prep_rel < LISZT : LST : TL : TL : TL : HD : EVENT > = < LISZT : LST : HD : EVENT > < LISZT : LST : TL : TL : TL : HD : HANDEL > = < LISZT : LST : HD : HANDEL > < LISZT : LST : TL : TL : TL : TL > = < LISZT : LAST > . ;;; Stuff after this point is just for test purposes reg-raised-np-sem (nom-obj) < LISZT : LST : HD > = name_rel < LISZT : LST : TL > = < LISZT : LAST > < HANDEL > = < LISZT : LST : HD : HANDEL > < INDEX > = < LISZT : LST : HD : INST >. p-agt-cause-transfer_rel (p-agt_rel). p-pat-recip_rel (prep_rel). name_rel (basic_nom_rel) < NAME > = string . give_rel (verb_rel). flap_rel (verb_rel). wave_rel (verb_rel). eat_rel (verb_rel). rabbit_rel (nom_rel) < INST : QUALIA > = animal. cheese_rel (nom_rel) < INST : QUALIA > = comestible.