************************************************
***Use of the IT-TB Tagset in the CoNLL files***
************************************************

In the IT-TB, each word is assigned 11 morphological codes.
If a code does not fit a word (e.g. "mood" does not fit nouns), the code "-" is assigned. The CoNLL files report information about the 'fitting' codes only (i.e. no "-").

In this document, "position" refers to the first column (P) of the IT-TB tagset file ("Tagset_IT.pdf": available at http://itreebank.marginalia.it/doc/Tagset_IT.pdf).
Each position corresponds to one attribute (second column of "Tagset_IT.pdf": ATT).
The codes permitted in each position (i.e. for each attribute) and their respective values are reported in the columns "C" (codes) and "VAL" (values) of "Tagset_IT.pdf".


A) Fine-grained PoS (column 5 in CoNLL files)
A label formed by two codes:
- Flexional category: position 3
- Flexional type: position 1

EXAMPLE: fine-grained PoS "B1"
- flexional category: II decl. (B)
- flexional type: nominal (1)

NOTES:
- column 4 in CoNLL files includes the coarse-grained PoS, which corresponds to the code for Flexional Type
Example: Coarse-grained PoS "1" -> flexional type: nominal
- Punctuations are labelled with PoS "Punc"
- There are few cases of PoS labels formed by one code (instead of two). In these cases, the code is a "flexional type" one (position 1).
Example: fine-grained PoS "5" (equal to coarse-grained PoS) -> flexional type: pseudo-lemma


B) Morph (column 6 in CoNLL files)
"Morph" features one or more values separated by |.
Values are formed by the "attribute name" (second column in "Tagset_IT.pdf": ATT) + "code" (fourth column in "Tagset_IT.pdf": C).
The attribute names are abbreviated as follows (three-letters abbreviations):
grn: position 2
mod: position 4
tem: position 5
grp: position 6
cas: position 7
gen: position 8
com: position 9
var: position 10
vgr: position 11

EXAMPLE: "grn1|casA|gen3":
- nominal degree ("grn"): positive (1)
- case/number ("cas"): singular nominative (A) 
- gender ("gen"): neuter (3) 

NOTES:
- there are words that are not assigned any morphological value (e.g. prepositions and conjunctions). In these cases, the "morph" column is filled with an underscore ("_")
- solution of the abbreviations for the attribute names:
grn: NomDegree
mod: Mood
tem: Tense
grp: PartDegree
cas: CaseNum
gen: GendNumPers
com: Comp
var: FVar
vgr: GVar