# Divvun & Giellatekno - open source grammars for Sámi and other languages
# Copyright © 2000-2010 The University of Tromsø & the Norwegian Sámi Parliament
# http://giellatekno.uit.no & http://divvun.no
#
# This program is free software; you can redistribute and/or modify
# this file under the terms of the GNU General Public License as published by
# the Free Software Foundation, either version 3 of the License, or
# (at your option) any later version. The GNU General Public License
# is found at http://www.gnu.org/licenses/gpl.html. It is
# also available in the file $GTHOME/LICENSE.txt.
#
# Other licensing options are available upon request, please contact
# giellatekno@hum.uit.no or divvun@hum.uit.no
# ==================================================================== #
# N O R T H E R N S Á M I D I S A M B I G U A T O R
# ==================================================================== #
# ====================================================================
# Sámi language technology project 2003-2006, University of Tromsø #
# From 22.02.08 this is the only up-to-date dis file. (Lene)
# Newest rule is r3589.
# =========================================== #
# Short table of contents #
# =========================================== #
# Delimiters
# Tags and sets
# Disambiguation rules
# One-cohort disambiguation - cycle 0
# Local disambiguation - cycles 1 and 2
# Cycle 1a: Special cases
# Cycle 1b: Cleaning up after the special cases
# Cycle 2: Other local disambiguation
# Mapping of COMP-CS<, CC and CS
# More diambiguation
# Verb mappings
# Disambiguating nouns
# Mainly mapping-rules
# Disambiguating nouns
# Case disambiguation
# Cycle 3: Global disambiguation
# Cycle 4: Syntactic disambiguation
# Cycle 5: Post-syntactic morphological disambiguation
# =========================================== #
# Detailed table of contents #
# =========================================== #
# Delimiters
# Tags and sets
# Tags declared as single-membered LISTs
# Parts of speech
# POS sub-categories
# Morphosyntactic properties
# Derivation
# Syntactic tags
# Semantic tags
# Other tags
# Single-word sets
# Initials
# Sets
# Word or not
# Derivational affixes
# Case sets
# Verb sets
# Pronoun sets
# Adjectival sets and their complements
# Coordinators
# Sets for adverbs that have lookalikes
# Sets of elements with common syntactic behaviour
# Verbs
# Adverbs
# Adjectives
# NP sets defined according to their morphosyntactic features
# The PRE-NP-HEAD family of sets
# Other negatively defined morphosyntactic noun sets
# Nominal sets defined according to their morphophonological properties
# Sets for lexeme homonymy
# Nominal sets defined according to their semantical properties
# Nouns with morpho-syntactic preferences
# Set hierarchy
# Miscellaneous sets
# Border sets and their complements
# Syntactic sets
# Disambiguation rules
# One-cohort disambiguation - cycle 0
# The section for popular rules
# Proper nouns
# Lexicalised derivations
# Acronyms
# Lexicalised adjectives
# Particular adjectives
# Lexicalised adverbs
# Adjectives or nouns, not adverbs
# Adverbs, not nouns or adjectives
# N Ess vs. A Ess
# Nouns, not verbs
# Nouns, verbs, or numerals
# Particular nouns
# Not noun
# Temporal expressions with locative case
# Southern Locative vs. Essive
# Px
# Adposition + Px
# Some verb rules
# Particular verbs
# Vowel-lowering
# Diphtong or not (empty)
# Compounds
# Local disambiguation - cycles 1 and 2
# Cycle 1a: cases that will be needed in careful mode later
# FAMILY pronouns
# An early rule for "eanaš"/"eanas"
# Px constraints
# Certain verb readings
# Cycle 1b: removing the readings that were left from the 1a rules
# Removing (nearly) all remaining Px readings
# Cycle 2: Other cases of local disambiguation
# Noun or not?
# Based on left context
# Based on right context
# Based on circumscribed context
# Some particular nouns
# Early possessive rule
# Adjectives and adverbs
# Disambiguationg abbreviations
# Case disambiguation of abbreviations
# Disambiguating particles
# Disambiguating clitics
# Disambiguating numerals
# Disambiguating adpositions
# Rules relating to particular adpositions
# Adpositions that take modifiers
# The case of GASKAL
# Po or Pr?
# Removing some readings that compete with Po
# Disambiguating subjunctions
# Some particular subjunctions
# CS selection for specific CS candidates
# CS removal for specific CS candidates
# Some more particular subjunctions
# Disambiguating adverbs
# General adverb rules
# Adverbs, not interjections
# Adjectives, not adverbs
# Adverbs, not comparative or plural adjectives
# Adverbs, not verbs
# Verbs, not adverbs
# Adverbs, not pronouns
# Adverbs, not nouns
# Nouns, not adverbs
# Rules related to specific adverbs (quite a lot)
# Mapping of COMP-CS<, CC and CS
# Complements of particles in comparison
# CC- and CS-Mapping
# More disambiguation
# Disambiguating pronouns
# Interrogative and relative pronouns
# Reflexive pronouns
# Reciprocal pronouns
# Numerals
# Numerals or not
# General rules
# Rules related to specific numerals
# More pronouns
# Indefinite pronouns
# General rules
# Particular indefinite pronouns
# Demonstrative pronouns
# DP-internal number concord
# DP-internal case concord
# FAMILY pronouns
# Some rules for FAMILY pronouns again (will hit "dat", except the last one)
# Dem or Pers
# Number
# Disambiguating against other POS
# Disambiguating adjectives
# Rules related to specific adjectives
# Comparative
# Attribute disambiguation
# Only Attr if something to lean on
# Rules for Attr between Dem and N
# Other attribute rules
# Coordinated attributes
# Special rules for 'buorre' (the only adjective showing case agreement)
# Comparatives
# Adjectives as predicates
# Coordination
# And now some rules for adverbs that modify adjectives
# Proper nouns
# SubstitutingProp tags
# Proper nouns
# Prop or not
# Particular proper nouns
# Prop as Attr or not
# Choosing between different proper nouns
# Disambiguating verbs - part 1
# ConNeg forms
# Imperative
# Infinitive
# Verbgenitive
# Supinum vs. potential
# Perfect Participle
# Der/NomAct og Actio
# Present participle
# Rules for "orrut"
# Rules for "addit" (which is an adjective, but more often a verb)
# Der/NomAct Loc = Der/NomAct Ess
# Actio Gen = Actio Ess
# Imprt or Ind
# Nouns vs. verbs
# Verb mappings
# Disambiguating nouns
# Mainly mapping-rules
# Lexically determined mappings
# Mappings of minor categories
# Determiner mappings
# Attribute mappings
# Numeral mappings
# Quantifier mappings
# Genitive mappings
# Appositions
# Fragment mappings
# Other special nominative mappings
# Arguments and secondary predicates
# Habitative mappings
# Adverbial mappings
# Disambiguating proper nouns #
# Choosing between different proper nouns
# Disambiguating the semantic tags
# Disambiguating symbols and punctuation marks
# Case disambiguation
# Secure rules for choosing Acc
# Choosing accusative or genitive semantically
# Genitive
# Nominative and accusative
# Nominative
# Accusative
# Accusative Px or locative
# Illative
# Early comitative rules
# Locative
# Comitative
# Locative and comitative - Disambiguation based on verbs
# Comitative
# Locative and comitative - Disambiguation based upon coordination
# Another round of locative rules
# Essive
# Late case rules (after other case rules have worked)
# Disambiguating verbs - part 2
# Finite or not
# Indicative or imperative
# Verbs according to person and number
# Rules for a special infinitive construction
# More finite verbs
# Passive
# Non-finite verb forms
# Infinitive
# Present Participle
# Actio/Perfect Participle
# Imperative II (empty)
# Verbgenitive
# Derived vs. lexicalized noun
# Aktio
# Selecting some more finite verbs
# Lexical disambiguation of verbs
# Residual cases
# Case rules
# Other rules for nouns and pronouns
# Px rules
# Determiners
# Adverbs and adjectives
# Cycle 3: Global disambiguation
# Nouns
# Verbs
# Genitive after preposition -- round 2
# Cycle 4: Syntactic disambiguation
# NP internal
# Verbal functions
# Subject
# Subject predicate
# Subject or not
# Late nominative rule
# Fragments
# Object and object predicate
# Adverbials or Object
# Other grammatical functions
# Cycle 5: Post-syntactic morphological disambiguation
# Demonstratives
# Case
# Adverbs
# Verbs
# Tidying up some loose ends
# Substituting some syntactic tags
# ================================================================= #
# ================================================================= #
# ================================================================= #
# XXX Copy to sme-dep.rle starts here.
# ========== #
# Delimiters #
# ========== #
DELIMITERS = "<.>" "" ">" "<...>" "<¶>";
#PARENTHESES = "<$(>" "<$)>" "<$[>" "<$]>" "<${>" "<$}>" "<$«>" "<$»>" ;
# ============= #
# Tags and sets #
# ============= #
SETS
# I define end of clause and beginning of clause in a way so that the file
# may be read both by the CG-2 and the vislcg formalisms.
LIST BOS = (>>>) ();
LIST EOS = (<<<) ();
# CG3 doesn´t function without >>> and <<< !
# Tags declared as single-membered LISTs
# ======================================
# Parts of speech
# ---------------
LIST N = N ;
LIST A = A ;
LIST Adv = Adv ;
LIST V = V ;
LIST Pron = Pron ;
LIST CS = CS ;
LIST CCC = CC ;
SET CC = CCC - ("sihke") ;
SET CC-CS = CC OR CS ;
LIST Po = Po ;
LIST Pr = Pr ;
LIST Pcle = Pcle ;
LIST Num = Num ;
LIST Coll = Coll ;
LIST NomAg = NomAg ;
LIST G3 = G3 ;
LIST Prop = Prop ;
LIST LEFT = LEFT ;
LIST RIGHT = RIGHT ;
LIST WEB = WEB ;
LIST V* = V* ;
# POS sub-categories
# ------------------
LIST Pers = Pers ;
LIST Dem = Dem ;
LIST Interr = Interr ;
LIST Indef = Indef ;
LIST Recipr = Recipr ;
LIST Refl = Refl ;
LIST Rel = Rel ;
LIST Interj = Interj ;
LIST Adp = Adp ;
# Morphosyntactic properties
# --------------------------
LIST Nom = Nom ;
LIST Acc = Acc ;
LIST Gen = Gen ;
LIST Ill = Ill ;
LIST Loc = Loc ;
LIST Com = Com ;
LIST Ess = Ess ;
LIST Sg = Sg ;
LIST Pl = Pl ;
LIST Du = Du ;
LIST RCmpnd = RCmpnd ;
LIST Cmpnd = Cmpnd ;
LIST SgNomCmp = SgNomCmp ;
LIST SgGenCmp = SgGenCmp ;
LIST ShCmp = ShCmp ;
LIST PxSg1 = PxSg1 ;
LIST PxSg2 = PxSg2 ;
LIST PxSg3 = PxSg3 ;
LIST PxDu1 = PxDu1 ;
LIST PxDu2 = PxDu2 ;
LIST PxDu3 = PxDu3 ;
LIST PxPl1 = PxPl1 ;
LIST PxPl2 = PxPl2 ;
LIST PxPl3 = PxPl3 ;
LIST Comp = Comp ;
LIST Superl = Superl ;
LIST Attr = Attr ;
LIST Ord = Ord ;
LIST Qst = Qst ;
# The ("ge" Pcle) etc. are used in Apertium
LIST Foc/ge = Foc/ge ("ge" Pcle) ;
LIST Foc/gen = Foc/gen ("gen" Pcle) ;
LIST Foc/ges = Foc/ges ("ges" Pcle) ;
LIST Foc/gis = Foc/gis ("gis" Pcle) ;
LIST Foc/naj = Foc/naj ("naj" Pcle) ;
LIST Foc/ba = Foc/ba ("ba" Pcle) ;
LIST Foc/be = Foc/be ("be" Pcle) ;
LIST Foc/hal = Foc/hal ("hal" Pcle) ;
LIST Foc/han = Foc/han ("han" Pcle) ;
LIST Foc/bat = Foc/bat ("bat" Pcle) ;
LIST Foc/son = Foc/son ("son" Pcle) ;
LIST IV = IV ;
LIST TV = TV ;
LIST Der/Pass = Der/PassL Der/PassS ;
LIST Prt = Prt;
LIST Prs = Prs ;
LIST Ind = Ind ;
LIST Pot = Pot ;
LIST Cond = Cond ;
LIST Imprt = Imprt ;
LIST ImprtII = ImprtII ;
LIST Sg1 = Sg1 ;
LIST Sg2 = Sg2 ;
LIST Sg3 = Sg3 ;
LIST Du1 = Du1 ;
LIST Du2 = Du2 ;
LIST Du3 = Du3 ;
LIST Pl1 = Pl1 ;
LIST Pl2 = Pl2 ;
LIST Pl3 = Pl3 ;
LIST Inf = Inf ;
LIST ConNeg = ConNeg ;
LIST Neg = Neg ;
LIST PrfPrc = PrfPrc ;
LIST VGen = VGen ;
LIST PrsPrc = PrsPrc ;
LIST Ger = Ger ;
LIST Sup = Sup ;
LIST Actio = Actio ;
LIST Der/NomAg = Der/NomAg ;
LIST Actor = NomAg Der/NomAg ;
LIST VAbess = VAbess ;
# Derivation
# ----------
LIST Der/adda = Der/adda ;
LIST Der/ahtti = Der/ahtti ;
LIST Der/alla = Der/alla ;
LIST Der/asti = Der/asti ;
LIST Der/easti = Der/easti ;
LIST Der/d = Der/d ;
LIST Der/eaddji = Der/eaddji ;
LIST Der/eamoš = Der/eamoš ;
LIST Der/amoš = Der/amoš ;
LIST Der/geahtes = Der/geahtes ;
LIST Der/gielat = Der/gielat ;
LIST !better: = !better: ;
LIST Der/NuA = Der/NuA ;
LIST Der/h = Der/h ;
LIST Der/heapmi = Der/heapmi ;
LIST Der/hudda = Der/hudda ;
LIST Der/huhtti = Der/huhtti ;
LIST Der/huvva = Der/huvva ;
LIST Der/halla = Der/halla ;
LIST Der/j = Der/j ;
LIST Der/l = Der/l ;
LIST Der/laš = Der/laš ;
LIST Der/las = Der/las ;
LIST Der/hat = Der/hat ;
LIST Der/meahttun = Der/meahttun ;
LIST Der/muš = Der/muš ;
LIST Der/NomAct = Der/NomAct ;
LIST Der/st = Der/st ;
LIST Der/stuvva = Der/stuvva ;
LIST Der/upmi = Der/upmi ;
LIST Der/supmi = Der/supmi ;
LIST Der/vuohta = Der/vuohta ;
LIST Der/goahti = Der/goahti ;
LIST Der/lágan = Der/lágan ;
LIST Der/lágán = Der/lágán ;
LIST Der/lágaš = Der/lágaš ;
LIST Der/jagáš = Der/jagáš ;
LIST Der/jahkásaš = Der/jahkásaš ;
LIST Der/diibmosaš = Der/diibmosaš ;
LIST Der/dássásaš = Der/dássásaš ;
LIST Der/Dimin = Der/Dimin ;
LIST Der/viđá = Der/viđá ;
LIST Der/viđi = Der/viđi ;
LIST Der/veara = Der/veara ;
LIST Der/duohke = Der/duohke ;
LIST Der/duohkai = Der/duohkai ;
LIST Der/vuolle = Der/vuolle ;
LIST Der/vuollai = Der/vuollai ;
LIST Der/vuolde = Der/vuolde ;
# Syntactic tags
# --------------
LIST @+FAUXV = @+FAUXV ;
LIST @+FMAINV = @+FMAINV ;
LIST @-FAUXV = @-FAUXV ;
LIST @-FMAINV = @-FMAINV ;
LIST @-FSUBJ> = @-FSUBJ> ;
LIST @-F = @-FOBJ> ;
LIST @SPRED = @-FADVL> ;
LIST @-F = @-FSPRED> ;
LIST @-FOPRED> = @-FOPRED> ;
SET FOBJ = @-F ;
SET FMAINV = @-FMAINV OR @+FMAINV ;
SET FAUXV = @-FAUXV OR @+FAUXV ;
LIST @ADVL = @ADVL @ADVL> @ADVL< @ADVL @-F ;
LIST @>ADVL = @>ADVL ;
LIST @ADVL< = @ADVL< ;
LIST @ = @ADVL> ;
LIST @HAB = @HAB ;
LIST @>N = @>N ;
LIST @>Interj = @>Interj ;
LIST @N< = @N< ;
LIST @>A = @>A ;
LIST @P< = @P< ;
LIST @>P = @>P ;
LIST @HNOUN = @HNOUN ;
LIST @INTERJ = @INTERJ ;
LIST @>Num = @>Num;
LIST @Pron< = @Pron< ;
LIST @>Pron = @>Pron ;
LIST @Num< = @Num< ;
LIST @OBJ = @OBJ ;
LIST @ = @OBJ> ;
LIST @OPRED = @OPRED ;
LIST @ = @OPRED> ;
LIST @PCLE = @PCLE ;
LIST @COMP-CS< = @COMP-CS< ;
LIST @SPRED = @SPRED ;
LIST @ = @SPRED> ;
LIST @SUBJ = @SUBJ ;
LIST @ = @SUBJ> ;
SET SUBJ = @ OR @SUBJ ;
SET SPRED = @ OR @SPRED ;
SET OPRED = @ OR @OPRED ;
LIST @PPRED = @PPRED ;
LIST @APP = @APP ;
LIST @APP-N< = @APP-N< ;
LIST @APP-Pron< = @APP-Pron< ;
LIST @APP>Pron = @APP>Pron ;
LIST @APP-Num< = @APP-Num< ;
LIST @APP-ADVL< = @APP-ADVL< ;
LIST @VOC = @VOC ;
LIST @CVP = @CVP ;
LIST @CNP = @CNP ;
SET OBJ = (@) OR (@OBJ) OR (@-F) ;
LIST = @OBJ> @-FOBJ> ;
SET -OTHERS = OBJ> OR (Gen) OR (Nom) OR (Ess) OR (Loc) OR (Adv) ;
LIST @X = @X ;
# Semantic tags
# -------------
LIST Ani = Ani ;
LIST Fem = Fem ;
LIST Mal = Mal ;
LIST Obj = Obj ;
LIST Org = Org ;
LIST Plc = Plc ;
LIST Sur = Sur ;
#LIST Tit = Tit ;
LIST Lang = Lang ;
LIST Plant = Plant ;
LIST Group = Group ;
LIST Time = Time ;
LIST Txt = Txt ;
LIST Route = Route ;
LIST Measr = Measr ;
LIST Wthr = Wthr ;
LIST Build = Build ;
LIST Edu = Edu ;
LIST Veh = Veh ;
LIST Event = Event ;
# Valency tags
# ------------
LIST = ;
LIST <0> = <0> ;
LIST = ;
LIST = ;
LIST = ;
LIST = ;
LIST = ;
LIST = ;
LIST = ;
LIST = ;
LIST = ;
LIST = ;
LIST = ;
LIST = ;
LIST = ;
LIST = ;
LIST = ;
LIST = ;
LIST = ;
LIST = ;
LIST = ;
LIST = ;
LIST = ;
LIST = ;
LIST = ;
LIST = ;
LIST = ;
LIST = ;
LIST = ;
LIST = ;
LIST = ;
LIST = ;
LIST = ;
LIST = ;
LIST = ;
LIST = ;
LIST = ;
LIST = ;
LIST = ;
LIST = ;
LIST = ;
LIST = ;
LIST = ;
LIST = ;
LIST = ;
LIST = ;
LIST = ;
LIST = ;
LIST = ;
LIST = ;
LIST = ;
LIST = ;
LIST = ;
LIST = ;
LIST = ;
LIST = ;
LIST = ;
LIST = ;
LIST = ;
LIST = ;
LIST = ;
LIST = ;
LIST = ;
LIST = ;
LIST = ;
LIST = ;
LIST = ;
LIST = ;
LIST = ;
LIST = ;
LIST = ;
LIST = ;
LIST = ;
LIST = ;
LIST = ;
LIST = ;
LIST = ;
LIST = ;
LIST = ;
LIST = ;
LIST = ;
LIST = ;
LIST = ;
LIST = ;
LIST = ;
LIST = ;
LIST = ;
LIST