SYNTAGMA: a brief overview


1. SYNTAGMA is a language processing pipeline, which includes tokenizer, pos-tagger, parser, NER, semantic-labeling.

Text-tagging and sentiment-analysis modules are currently under development.


2. Source code is provisionally written in FreePascal. Lexical-, NE- and semantic data are stored as MySQL tables.


3. It is a deterministic, rule-based system, not a statistical one. There are some "probabilistic" rules, for morphological and syntactic disambiguation (they assign a score), but they are still rules, since my assumption is that human language competence doesn't work statistically.

Attachment 1) shows some parse results.


4. There are some fundamental data linked with lexical entries in order to make SYNTAGMA work properly. The most important are the following two:


4.1 For each lexical entry there is an index of at least one meaning (MNG), i.e. word-sense.

Meanings are linked with data related to their argument structure, that include information about subcategorisation, connectives and semantic properties of each argument.

These information are crucial for SYNTAGMA, since it works on a cross-verification between syntactic and semantic properties in order to solve structural (mostly dependency related) ambiguities, by select the appropriate pattern ('phrase-structure'), and to perform WSD tasks according to the possible syntactical interpretations (i.e. argument structures) of a sentence.

These data for Italian have been extracted from an Italian dictionary that includes information about the valence (Tesnière 1959) of lexical entries. For English Syntagma refers to the frames linked to the WordNet synsets.

Attachment 2) shows some argument frames of different categories of words.


4.2 The parser works bottom-up, building syntagmatic units based on a list of patterns. Patterns describe syntactic structures whose complexity gradually increases, from terminal elements to complex sentences, including dislocations and elliptical constructions and what else you want to include in syntagma's syntactic "knowledge".

The difference between a 'phrase structure' (as in Generative Grammar) and the 'pattern' is that this latter describes possible part of- and sentence structures, rather than 'abstract' constituents, but maintaining recursiveness.

This list of patterns ("phrase structures") is given in a text file, so it can be written and managed very easily.

Attachment 3) shows an extract of the Pattern_List (Italian).



Daniel Christen (www.semasia.ch)


SYNTAGMA. A Linguistic Approach to Parsing










ATTACHMENT 1. SYNTAGMA. Parse examples. Daniel Christen (www.semasia.ch). October 2019.



* Note: since 2023 the Syntagma parser works whitout traces; raised NPs and clitics are linked directly to argument-structure of the embedded verb.


EXAMPLE (1) input= "La bambina glielo vorrebbe poter regalare." ("The girl would like to be able to give it to him").

The example shows a sentence with two encapsulated modal verbs ("want" and "can") and two pro-clitic pronouns ("gli" Dative and "lo" Accusative) which are antecedents for the arguments of the verb "regalare" ("give") in the subordinate clause; these arguments are given by two Traces, with a coreference pointer (REF) to its antecedents.

Italian allows phonetically unexpressed subjects, which are represented by the symbol PRO (Null-subject) or Pro (non-finite clause subject), following Gouvernement and Bynding Theory.

The unexpressed subject Pro has a coreference pointer (REF) to the subject of the main-verb.

The coreference index (REF) refers to the terminal index (TRM) of the antecedent.


Note well that the parse selects, for each lexical entry, the word-sense/word-senses (MNG) that are compatible with the selected syntactic structure. The other word-senses are deleted from

the meaning-index array of the given lexical entry. This is the result of a FILTER that considers the relation between a meaning and its argument frame (listed in the argument frame table).


Parsing completed: s=1 global_id_snt=1 tot_words=7.

TEMA: snt=1 frasi[snt]="la bambina glielo vorrebbe poter regalare.". Selected result:


205;F1; snt=1;4101; SUCCESS=1; INHIBITED=FALSE; NUM_PAR=7; E_TYPE=ASSERT; S_TYPE=; HEAD=8; DOM=(203);(IMs=552467;volere; DISCs=86307; MDVs=2; TMPs=1; PRSs=3; GENs=0; NUMs=1;)

1;la; LEX: la; ID=1; TRM=1; RTRM=1; GOV=4; FNCT=det; TAGS=NP; CAT=613; |1-2|; MORF [(IM=150264; LEX=la; CAT=613; DISC=477; MDV=0; TMP=0; PRS=3; GEN=2; NUM=1; Lsem=0; RESTR_LEX= (1);)] BOUND=0; DIAT=; LogicSPACE -1

2;bambina; LEX: bambino; ID=2; TRM=4; RTRM=4; GOV=8; FNCT=sogg; TAGS=NP; CAT=200; |2-3|; MORF [(IM=29563; LEX=bambino; CAT=200; DISC=8110; MDV=0; TMP=0; PRS=3; GEN=2; NUM=1; Lsem=0; RESTR_LEX= (1);)] BOUND=0; DIAT=; MNG(8110_01) LogicSPACE -1

3;gli; LEX: gli; ID=3; TRM=6; RTRM=6; GOV=10; FNCT=rc; TAGS=PronDat; CAT=511; |3-4|; MORF [(IM=121527; LEX=gli; CAT=511; DISC=1278; MDV=0; TMP=0; PRS=3; GEN=0; NUM=0; Lsem=0; RESTR_LEX= (1);)] BOUND=462531; DIAT=; LogicSPACE -1

4;lo; LEX: lo; ID=3; TRM=7; RTRM=7; GOV=10; FNCT=ro; TAGS=PronAcc; CAT=512; |4-5|; MORF [(IM=552740; LEX=lo; CAT=512; DISC=1297; MDV=0; TMP=0; PRS=3; GEN=1; NUM=1; Lsem=0; RESTR_LEX= (1);)] BOUND=462531; DIAT=; LogicSPACE -1

5;vorrebbe; LEX: volere; ID=4; TRM=8; RTRM=8; GOV=0; FNCT=v; TAGS=VP; CAT=100; |5-6|; MORF [(IM=552467; LEX=volere; CAT=100; DISC=86307; MDV=2; TMP=1; PRS=3; GEN=0; NUM=1; Lsem=0; RESTR_LEX= (1);)] BOUND=0; DIAT=att; MNG(86307_51) LogicSPACE -1

6;PRO; LEX: PRO; ID=0; TRM=13; RTRM=13; GOV=10; FNCT=sogg; TAGS=NP; CAT=pro; |0-0|; MORF [(IM=0; LEX=PRO; CAT=; DISC=0; MDV=0; TMP=0; PRS=3; GEN=2; NUM=1; Lsem=; RESTR_LEX= (0);)(IM=0; LEX=; CAT=; DISC=; MDV=0; TMP=0; PRS=3; GEN=2; NUM=1; Lsem=; RESTR_LEX= (0);)] BOUND=0; DIAT=; LogicSPACE -1 REF(4:0)

7;potere; LEX: potere; ID=5; TRM=10; RTRM=10; GOV=8; FNCT=arg; TAGS=VP; CAT=100; |6-7|; MORF [(IM=215796; LEX=potere; CAT=100; DISC=59288; MDV=6; TMP=1; PRS=0; GEN=0; NUM=0; Lsem=0; RESTR_LEX= (1);)(IM=542214; LEX=potere; CAT=100; DISC=59288; MDV=6; TMP=1; PRS=0; GEN=0; NUM=0; Lsem=0; RESTR_LEX= (1);)] BOUND=0; DIAT=; MNG(59288_01) LogicSPACE -1

8;PRO; LEX: PRO; ID=0; TRM=13; RTRM=13; GOV=11; FNCT=sogg; TAGS=NP; CAT=pro; |0-0|; MORF [(IM=0; LEX=PRO; CAT=; DISC=0; MDV=0; TMP=0; PRS=0; GEN=0; NUM=0; Lsem=; RESTR_LEX= (0);)] BOUND=0; DIAT=; LogicSPACE -1 REF(13:0)

9;regalare; LEX: regalare; ID=6; TRM=11; RTRM=11; GOV=10; FNCT=arg; TAGS=VP; CAT=100; |7-8|; MORF [(IM=460630; LEX=regalare; CAT=100; DISC=63732; MDV=6; TMP=1; PRS=0; GEN=0; NUM=0; Lsem=0; RESTR_LEX= (1);)] BOUND=0; DIAT=; MNG(63732_01) LogicSPACE -1

10;T; LEX: T; ID=0; TRM=14; RTRM=14; GOV=11; FNCT=arg; TAGS=NP; CAT=trace; |0-0|; MORF [(IM=0; LEX=T; CAT=; DISC=0; MDV=0; TMP=0; PRS=3; GEN=1; NUM=1; Lsem=; RESTR_LEX= (0);)] BOUND=0; DIAT=; LogicSPACE -1 REF(7:0)

11;T; LEX: T; ID=0; TRM=14; RTRM=14; GOV=11; FNCT=prep.arg; TAGS=PP; CAT=trace; |0-0|; MORF [(IM=0; LEX=T; CAT=; DISC=0; MDV=0; TMP=0; PRS=3; GEN=0; NUM=0; Lsem=; RESTR_LEX= (0);)] BOUND=0; DIAT=; LogicSPACE -1 REF(6:0)

12;.; LEX: .; ID=7; TRM=12; RTRM=12; GOV=8; FNCT=sep; TAGS=Sep21; CAT=11; |8-9|; MORF [(IM=553196; LEX=.; CAT=11; DISC=1; MDV=0; TMP=0; PRS=0; GEN=0; NUM=0; Lsem=0; RESTR_LEX= (1);)] BOUND=0; DIAT=; LogicSPACE -1

SCORE= 0.00000000000000E+000; SEM_SCORE= 0.00000000000000E+000; TestWon/Done=0/0 Competitors=179,

History=()



EXAMPLE (2) input= "L'applicativo funziona bene ma è troppo caro." ("The app works well, but it is too expensive.")

Parse with sentiment analysis.

The unexpressed subject PRO of the coordinate clause has a coreference pointer (REF=4) to the terminal index (TRM) subject ("applicativo") of the first clause verb.


Parsing completed: s=1 global_id_snt=1 tot_parole=9.

TEMA: snt=1 frasi[snt]="L' applicativo funziona bene ma è troppo caro.". Selected result:

307;F1; snt=1;4101; SUCCESS=1; INHIBITED=FALSE; NUM_PAR=9; E_TYPE=ASSERT; S_TYPE="COORD"; HEAD=12; LOGIC(11 0F1 "BUT") DOM=(297);(IMs=538721;essere; DISCs=27895; MDVs=1; TMPs=1; PRSs=3; GENs=0; NUMs=1;)

1;l'; LEX: l'; ID=1; TRM=1; RTRM=1; GOV=4; FNCT=det; TAGS=NP; CAT=613; |1-2|; MORF [(IM=552735; LEX=l'; CAT=613; DISC=476; MDV=0; TMP=0; PRS=3; GEN=0; NUM=1; Lsem=0; RESTR_LEX= (1);)] BOUND=0; DIAT=; GOV_COORD 0 LSPACE -1

2;applicativo; LEX: applicativo; ID=2; TRM=4; RTRM=4; GOV=5; FNCT=sogg; TAGS=NP; CAT=200; |2-3|; MORF [(IM=18750; LEX=applicativo; CAT=200; DISC=5043; MDV=0; TMP=0; PRS=3; GEN=1; NUM=1; Lsem=0; RESTR_LEX= (1);)] BOUND=0; DIAT=; MNG(5043_02) GOV_COORD 0 LSPACE -1

3;funziona; LEX: funzionare; ID=3; TRM=5; RTRM=5; GOV=0; FNCT=v; TAGS=VP; CAT=100; |3-4|; MORF [(IM=538721; LEX=essere; CAT=100; DISC=27895; MDV=1; TMP=1; PRS=3; GEN=0; NUM=1; Lsem=0; RESTR_LEX= (1);)] BOUND=0; DIAT=att; MNG(32205_01) GOV_COORD 0 LSPACE 11

4;bene; LEX: bene; ID=4; TRM=7; RTRM=7; GOV=5; FNCT=exp; TAGS=AdvP; CAT=300; |4-5|; MORF [(IM=32531; LEX=bene; CAT=300; DISC=9017; MDV=0; TMP=0; PRS=0; GEN=0; NUM=0; Lsem=0; RESTR_LEX= (1);)] BOUND=0; DIAT=; MNG(9017_01) GOV_COORD 0 LSPACE -1

5;ma; LEX: ma; ID=5; TRM=11; RTRM=11; GOV=12; FNCT=ccoord; TAGS=CngCoord01; CAT=715; |5-6|; MORF [(IM=158897; LEX=ma; CAT=715; DISC=508; MDV=0; TMP=0; PRS=0; GEN=0; NUM=0; Lsem=BUT; RESTR_LEX= (1);)] BOUND=0; DIAT=; GOV_COORD 0 LSPACE -1

6;PRO; LEX: PRO; ID=0; TRM=22; RTRM=22; GOV=12; FNCT=sogg; TAGS=NP; CAT=pro; |0-0|; MORF [(IM=0; LEX=PRO; CAT=; DISC=0; MDV=0; TMP=0; PRS=3; GEN=0; NUM=1; Lsem=; RESTR_LEX= (0);)] BOUND=0; DIAT=0; GOV_COORD 0 LSPACE -1 REF(4:0)

7;è; LEX: essere; ID=6; TRM=12; RTRM=12; GOV=0; FNCT=v; TAGS=VP; CAT=100; |6-7|; MORF [(IM=538721; LEX=essere; CAT=100; DISC=27895; MDV=1; TMP=1; PRS=3; GEN=0; NUM=1; Lsem=0; RESTR_LEX= (1);)] BOUND=0; DIAT=att; MNG(27895_33) GOV_COORD 0 LSPACE 11

8;troppo; LEX: troppo; ID=7; TRM=14; RTRM=14; GOV=18; FNCT=exp; TAGS=AdjP; CAT=310; |7-8|; MORF [(IM=299706; LEX=troppo; CAT=310; DISC=364; MDV=0; TMP=0; PRS=0; GEN=0; NUM=0; Lsem=QUANTITY|PLUS; RESTR_LEX= (1);)] BOUND=0; DIAT=; GOV_COORD 0 LSPACE -1

9;caro; LEX: caro; ID=8; TRM=18; RTRM=18; GOV=12; FNCT=compl.pred; TAGS=AdjP; CAT=400; |8-9|; MORF [(IM=47780; LEX=caro; CAT=400; DISC=13557; MDV=0; TMP=0; PRS=0; GEN=1; NUM=1; Lsem=0; RESTR_LEX= (1);)] BOUND=0; DIAT=; MNG(13557_01) GOV_COORD 0 LSPACE -1

10;.; LEX: .; ID=9; TRM=21; RTRM=21; GOV=12; FNCT=sep; TAGS=Sep21; CAT=11; |9-10|; MORF [(IM=553196; LEX=.; CAT=11; DISC=1; MDV=0; TMP=0; PRS=0; GEN=0; NUM=0; Lsem=0; RESTR_LEX= (1);)] BOUND=0; DIAT=; GOV_COORD 5,12 LSPACE -1

SCORE= 1.00000000000000E+000; SEM_SCORE= 2.00000000000000E-001; TestWon/Done=7/7 Competitors=210,211,212,214,301,302,303,304,305,306,308,309,310,333,334,335,336,

History=[267 W D7 170][267 W D7 253][307 W D7 214][307 W C8 (caro) 301][307 W C8 (caro) 302][307 W C8 (caro) 303][307 W C8 (caro) 305]


registra_parse_result: sintagma.sntgm=290 has_highest_score(sintagma)=TRUE

Parser: tot_s=1 s=1 global_id_snt=1 frasi[s]=L' applicativo funziona bene ma è troppo caro."

TEMA: process completed. Input=L'applicativo funziona bene ma è troppo caro."

FINE registra_sentiment_text: global_id_text=0 text_domain= sentiment_text_result= 1.05416666666660E+001

Parsing completed. Result is in file=dataout/xout_parser.txt

Parser: Press ENTER for next sentence!











ATTACHMENT 2. SYNTAGMA argument frames: examples. Daniel Christen, October 2019


The following tables show the argument structure (and subcategorisation frame) of the different word senses (meanings 'mng') of a lexical entry ('lex').

These examples are somewhere redundant: they bring together data coming from different tables (columns 'lex', prf', idiom' are pleonastic).


For each lexical entry ("lex" indexed by "disc") and for each word sense ('mng'), its possible argument structures (subconstituents) are listed.

For each argument/subcontituent, indexed by 'ival', the table gives:

- its syntactic function 'fnct' (sogg=subject; arg=direct object; prep.arg = indirect object ...)

- for each 'fnct' the table describes its possible surface forms and properties ('iprp'), that is:

- the syntactic label 'vcat' (NP, PP, VP, F=Clause, Finf=infinitive clause)

- the selected connectives 'conn': i.e. prepositions, conjuncions

- restrictions on the lexical form ('vlex') of the contituent

- restrictions on its morphological features (verbal mood 'vmdv', tempus 'vtmp', person 'vprs', number, gender)

- semantic label 'vsem' (which can have a restrictive role)

- anaphora link to some other subcontituent 'vref'

- special restrictions (given in a logical form: they are mostly in the Pattern-List file)

Data coming from other tables:

- lexical entry ('lex')

- gloss ('prf')

- 'idiom': 1=the word sense belongs to an idiom (idiomatic locution).


EXAMPLE 1: argument frame of "amare" (verb "love"): word senses (meanings=column 'mng') of the lexical entry (which is indexed as 'disc').

# id_frame, disc, mng, ival, fnct, opt, iprp, vcat, conn, vlex, vmdv, vtmp, vprs, vgen, vnum, vpos, vsem, vref, vrestr, lex, prf, idiom, verif

'1970', '3085', '3085_01', '1', 'sogg', 'F', '1', 'NP', '0', '0', '0', '0', '0', '0', '0', '0', '0', '0', '0', 'amare', 'Provare amore e affetto verso qlcu.', '0', '0'

'1970', '3085', '3085_01', '2', 'v', 'F', '1', 'VP', '0', '0', '0', '0', '0', '0', '0', '0', '0', '0', '0', 'amare', 'Provare amore e affetto verso qlcu.', '0', '0'

'1970', '3085', '3085_01', '3', 'arg', 'F', '1', 'NP', '0', '0', '0', '0', '0', '0', '0', '0', '0', '0', '0', 'amare', 'Provare amore e affetto verso qlcu.', '0', '0'


'1972', '3085', '3085_03', '1', 'sogg', 'F', '1', 'NP', '0', '0', '0', '0', '0', '0', '0', '0', '0', '0', '0', 'amare', 'essere innamorato di qlcu.', '0', '0'

'1972', '3085', '3085_03', '2', 'v', 'F', '1', 'VP', '0', '0', '0', '0', '0', '0', '0', '0', '0', '0', '0', 'amare', 'essere innamorato di qlcu.', '0', '0'

'1972', '3085', '3085_03', '3', 'arg', 'F', '1', 'NP', '0', '0', '0', '0', '0', '0', '0', '0', '0', '0', '0', 'amare', 'essere innamorato di qlcu.', '0', '0'


'1977', '3085', '3085_08', '1', 'sogg', 'F', '1', 'NP', '0', '0', '0', '0', '0', '0', '0', '0', '0', '0', '0', 'amare', 'provare diletto nel fare qlco. o desiderare, volere qlco.', '0', '0'

'1977', '3085', '3085_08', '2', 'v', 'F', '1', 'VP', '0', '0', '0', '0', '0', '0', '0', '0', '0', '0', '0', 'amare', 'provare diletto nel fare qlco. o desiderare, volere qlco.', '0', '0'

'1977', '3085', '3085_08', '3', 'arg', 'F', '1', 'Finf', 'no_conn|di', '0', '6', '0', '0', '0', '0', '0', 'THING', '0', '0', 'amare', 'provare diletto nel fare qlco. o desiderare, volere qlco.', '0', '0'

'1977', '3085', '3085_08', '3', 'arg', 'F', '2', 'F', 'che', '0', '1|2|3', '0', '0', '0', '0', '0', 'THING', '0', '0', 'amare', 'provare diletto nel fare qlco. o desiderare, volere qlco.', '0', '0'


'1978', '3085', '3085_09', '1', 'sogg', 'F', '1', 'NP', '0', '0', '0', '0', '0', '0', '0', '0', 'animal', '0', '0', 'amare', 'prediligere qlco.', '0', '0'

'1978', '3085', '3085_09', '2', 'v', 'F', '1', 'VP', '0', '0', '0', '0', '0', '0', '0', '0', '0', '0', '0', 'amare', 'prediligere qlco.', '0', '0'

'1978', '3085', '3085_09', '3', 'arg', 'F', '1', 'NP', '0', '0', '0', '0', '0', '0', '0', '0', 'THING', '0', '0', 'amare', 'prediligere qlco.', '0', '0'


'1979', '3085', '3085_10', '1', 'sogg', 'F', '1', 'NP', '0', '0', '0', '0', '0', '0', '0', '0', 'vegetal&plant', '0', '0', 'amare', 'abbisognare di qlco. per prosperare', '0', '0'

'1979', '3085', '3085_10', '2', 'v', 'F', '1', 'VP', '0', '0', '0', '0', '0', '0', '0', '0', '0', '0', '0', 'amare', 'abbisognare di qlco. per prosperare', '0', '0'

'1979', '3085', '3085_10', '3', 'arg', 'F', '1', 'NP', '0', '0', '0', '0', '0', '0', '0', '0', '0', '0', '0', 'amare', 'abbisognare di qlco. per prosperare', '0', '0'


EXAMPLE 2: argument frame of "decisione" (noun: "decision")

# id_frame, disc, mng, ival, fnct, opt, iprp, vcat, conn, vlex, vmdv, vtmp, vprs, vgen, vnum, vpos, vsem, vref, vrestr, lex, prf, idiom, verif

'0', '21481', '21481_01', '1', 'spc', 'T', '1', 'PP', 'di', '0', '0', '0', '0', '0', '0', '0', '0', '0', 'source: EXE (conn)', 'decisione', 'rimettersi alle decisione di qlcu.', '0', '1'

'0', '21481', '21481_03', '1', 'spc', 'T', '1', 'PP', 'di', '0', '0', '0', '0', '0', '0', '0', '0', '0', 'source: EXE (conn)', 'decisione', 'decisione del tribunale', '0', '1'


EXAMPLE 3: argument frame of "fedele" (adj: "faithful", "loyal")

'11200', '29318', '29318_03', '1', 'spc', 'F', '1', 'PP', 'a', '0', '0', '0', '0', '0', '0', '0', '0', '0', '0', 'fedele', 'che è costante nell\'amore, negli affetti, nelle proprie convinzioni', '0', '0'

'0', '29318', '29318_05', '1', 'spc', 'T', '1', 'PP', 'a', '0', '0', '0', '0', '0', '0', '0', '0', '0', 'source: EXE (conn)', 'fedele', 'fedele alla parola data, agli amici', '0', '1'

'0', '29318', '29318_06', '1', 'spc', 'T', '1', 'PP', 'a', '0', '0', '0', '0', '0', '0', '0', '0', '0', 'source: PRF (conn)', 'fedele', 'Conforme all\'originale o al vero (anche con la prep. a)', '0', '1'

'0', '29318', '29318_08', '1', 'spc', 'T', '1', 'PP', 'di', '0', '0', '0', '0', '0', '0', '0', '0', '0', 'source: EXE (conn)', 'fedele', 'fare un quadro fedele della situazione', '0', '1'

'0', '29318', '29318_12', '1', 'spc', 'T', '1', 'PP', 'di', '0', '0', '0', '0', '0', '0', '0', '0', '0', 'source: EXE (conn)', 'fedele', 'i fedele del re, del Milan', '0', '1'


EXAMPLE 4: some argument frames of "vedere" (verb: "see")

'33277', '84807', '84807_01', '1', 'sogg', 'F', '1', 'NP', '0', '0', '0', '0', '0', '0', '0', '0', 'ANIMATE|PERSON|ANIMAL', '0', '0', 'vedere', 'Percepire qlcu. o qlco. mediante la facoltà della vista', '0', '0'

'33277', '84807', '84807_01', '2', 'v', 'F', '1', 'VP', '0', '0', '0', '0', '0', '0', '0', '0', '0', '0', '0', 'vedere', 'Percepire qlcu. o qlco. mediante la facoltà della vista', '0', '0'

'33277', '84807', '84807_01', '3', 'arg', 'F', '1', 'NP', '0', '0', '0', '0', '0', '0', '0', '0', 'PERCEPTION|THING|PERSON|EVENT', '0', '0', 'vedere', 'Percepire qlcu. o qlco. mediante la facoltà della vista', '0', '0'


'33296', '84807', '84807_23', '3', 'arg', 'T', '2', 'Finterr', 'chi|che cosa|cosa|quando|perché', '0', '1|2|3', '0', '0', '0', '0', '0', 'EVENT', '0', '0', 'vedere', 'Prendere in esame, osservare qlco.', '0', '1'

'33297', '84807', '84807_24', '1', 'sogg', 'F', '1', 'NP', '0', '0', '0', '0', '0', '0', '0', '0', 'PERSON', '0', '0', 'vedere', 'considerare, controllare qlco.', '0', '0'

'33297', '84807', '84807_24', '2', 'v', 'F', '1', 'VP', '0', '0', '0', '0', '0', '0', '0', '0', '0', '0', '0', 'vedere', 'considerare, controllare qlco.', '0', '0'

'33297', '84807', '84807_24', '3', 'arg', 'T', '1', 'NP', '0', '0', '1|2|3', '0', '0', '0', '0', '0', 'THING', '0', '0', 'vedere', 'considerare, controllare qlco.', '0', '0'


'33307', '84807', '84807_34', '1', 'sogg', 'F', '1', 'NP', '0', '0', '0', '0', '0', '0', '0', '0', '0', '0', '0', 'vedere', 'seguita da frase, esprime atteggiamento di cautela, di sospensione del giudizio o anche di sfida e richiesta di verifica', '1', '0'

'33307', '84807', '84807_34', '2', 'v', 'F', '1', 'VP', '0', '0', '0', '0', '0', '0', '0', '0', '0', '0', '0', 'vedere', 'seguita da frase, esprime atteggiamento di cautela, di sospensione del giudizio o anche di sfida e richiesta di verifica', '1', '0'

'33307', '84807', '84807_34', '3', 'arg', 'F', '1', 'NP', '0', '0', '0', '0', '0', '0', '0', '0', '0', '0', '0', 'vedere', 'seguita da frase, esprime atteggiamento di cautela, di sospensione del giudizio o anche di sfida e richiesta di verifica', '1', '0'


'33318', '84807', '84807_52', '1', 'sogg', 'F', '1', 'NP', '0', '0', '0', '0', '0', '0', '0', '0', '0', '0', '0', 'vedere', 'Visitare, guardare qlco.', '0', '0'

'33318', '84807', '84807_52', '2', 'v', 'F', '1', 'VP', '0', '0', '0', '0', '0', '0', '0', '0', '0', '0', '0', 'vedere', 'Visitare, guardare qlco.', '0', '0'

'33318', '84807', '84807_52', '3', 'arg', 'F', '1', 'NP', '0', '0', '0', '0', '0', '0', '0', '0', 'THING', '0', '0', 'vedere', 'Visitare, guardare qlco.', '0', '0'

'33319', '84807', '84807_53', '1', 'sogg', 'F', '1', 'NP', '0', '0', '0', '0', '0', '0', '0', '0', '0', '0', '0', 'vedere', 'assistere a qlco.', '0', '0'

'33319', '84807', '84807_53', '2', 'v', 'F', '1', 'VP', '0', '0', '0', '0', '0', '0', '0', '0', '0', '0', '0', 'vedere', 'assistere a qlco.', '0', '0'

'33319', '84807', '84807_53', '3', 'arg', 'F', '1', 'NP', '0', '0', '0', '0', '0', '0', '0', '0', '0', '0', '0', 'vedere', 'assistere a qlco.', '0', '0'

'33320', '84807', '84807_54', '1', 'sogg', 'F', '1', 'NP', '0', '0', '0', '0', '0', '0', '0', '0', '0', '0', '0', 'vedere', 'Incontrare qlcu.', '0', '0'

'33320', '84807', '84807_54', '2', 'v', 'F', '1', 'VP', '0', '0', '0', '0', '0', '0', '0', '0', '0', '0', '0', 'vedere', 'Incontrare qlcu.', '0', '0'

'33320', '84807', '84807_54', '3', 'arg', 'F', '1', 'NP', '0', '0', '0', '0', '0', '0', '0', '0', 'PERSON', '0', '0', 'vedere', 'Incontrare qlcu.', '0', '0'