...

Nessun titolo diapositiva - ittig

by user

on
Category:

linguistics

23

views

Report

Comments

Transcript

Nessun titolo diapositiva - ittig
CONSIGLIO NAZIONALE DELLE RICERCHE
Istituto di Teoria e Tecniche dell’Informazione Giuridica
Formal Models for a
Legislative Grammar.
Explicit Text Amendment
Andrea Bolioli, Pietro Mercatali, Francesco Romano
KMGov 2004
Krems, 19 may 2004
CONSIGLIO NAZIONALE DELLE RICERCHE
Istituto di Teoria e Tecniche dell’Informazione Giuridica
http://www.ittig.cnr.it
Scope and Assumptions 1
For the communication of legislative sources
through the Internet, the parliamentary
and governmental institutions of many
countries have begun a process of
converting their “deposits” of these, into a
standard format
Krems 19 may 2004
2
CONSIGLIO NAZIONALE DELLE RICERCHE
Istituto di Teoria e Tecniche dell’Informazione Giuridica
http://www.ittig.cnr.it
Scope and Assumptions 2
• The XML mark-up language seems to be the tool
deputised for reaching this scope.
• This language combining its dual nature as a
mark-up language and a Web standard, is able
to form the common ground for action both “at
the source”, namely, legislative drafting, and
action “downstream” relating to the publication
of the texts and the identification of tools for
accessing legislative information
Krems 19 may 2004
3
CONSIGLIO NAZIONALE DELLE RICERCHE
Istituto di Teoria e Tecniche dell’Informazione Giuridica
http://www.ittig.cnr.it
Scope and Assumptions 3
• In order to adopt this language as a standard and,
above all, for the conversion of the legislative
instruments in force into the format provided for,
by the DTD rules, two factors, in our opinion,
must interact:
• A) Definition and promotion of a “controlled”
legislative language
• B) Use of tools for natural language recognition
Krems 19 may 2004
4
CONSIGLIO NAZIONALE DELLE RICERCHE
Istituto di Teoria e Tecniche dell’Informazione Giuridica
http://www.ittig.cnr.it
The Method
For this research, the methodological approach can be
subdivided into the following steps:
• identification of the technical tool for the implementation
of the parser and the extraction of the information;
• identification and description of the models, on the basis
of legal rules;
• identification and description of the textual structures
expressing the defined legal models;
• choice of the sample of legislative instruments to be
analysed, compilation of the grammar according to the
syntax of the pre-selected parser and the automated
analysis of the sample.
Krems 19 may 2004
5
CONSIGLIO NAZIONALE DELLE RICERCHE
Istituto di Teoria e Tecniche dell’Informazione Giuridica
http://www.ittig.cnr.it
Technical Tool for the Implementation of the
Parser and for Information Extraction (1)
The suitable tool for the recognition and tagging of a legislative
instrument, has been identified in the Sophia 2.1 system of parsing.
In particular, we are working with this software on analysing and
tagging the first sample of legislative instruments, in the following
phases:
• normalisation of the entry text, properly tagging all those structures
and textual segments that can be recognised on the basis of
characters or, in other words, without resort to or consultation of the
lexicon-dictionary;
• lexical (syntactical category) and morphological (flexion passages)
analysis of the text in input;
Krems 19 may 2004
6
CONSIGLIO NAZIONALE DELLE RICERCHE
Istituto di Teoria e Tecniche dell’Informazione Giuridica
http://www.ittig.cnr.it
Technical Tool for the Implementation of the
Parser and for Information Extraction (2)
• disambiguation of the syntactical category of the words
(Part of Speech Tagging);
• partial syntactical analysis (called chunking), aimed at
identifying the minimum syntactical groups present in the
text in input and at grouping them in constituents;
• semantic analysis and identification of the relevant
conceptual structures in the text in input;
• conversion of the analysed document from the original
format (Microsoft Word, HTML, RTF, txt, etc.) into the
XML format, according to the established DTD.
Krems 19 may 2004
7
CONSIGLIO NAZIONALE DELLE RICERCHE
Istituto di Teoria e Tecniche dell’Informazione Giuridica
http://www.ittig.cnr.it
Architecture
Syntactical
analysis
Semantic
analysis
Tokenization
and
XSL
morphology
Tagger XML
Docs
DB
Krems 19 may 2004
8
CONSIGLIO NAZIONALE DELLE RICERCHE
Istituto di Teoria e Tecniche dell’Informazione Giuridica
http://www.ittig.cnr.it
Identification and Description of the
Models on the Basis of Legal Rules
The legislative instrument has, by definition, a
prescriptive function, and in virtue of this, the request is
that the legislative instrument responds to a set of rules
that dominate and, at the same time, stand beside,
integrate, and sometimes modify the rules that make up
common language and these rules too, can be defined as
legal rules
For the implementation of the grammar that will be
utilised by the parser, it is necessary to integrate the
models extracted from the legal rules with the linguistic
rules. We call these models “well-formed”.
Krems 19 may 2004
9
CONSIGLIO NAZIONALE DELLE RICERCHE
Istituto di Teoria e Tecniche dell’Informazione Giuridica
http://www.ittig.cnr.it
Legislative delegation.
The delegation provision must comply:
• it must be conferred on the Government;
• it must contain a term within which the Government has
to enact the delegated act;
• it must specify the object of the delegation;
• it must contain the guiding principles and criteria to
which the Government has to adhere in the exercise of
that delegation.
Krems 19 may 2004
10
CONSIGLIO NAZIONALE DELLE RICERCHE
Istituto di Teoria e Tecniche dell’Informazione Giuridica
http://www.ittig.cnr.it
Tags or qualifiers of the elements making up
the structure
<ADDRESSEE> The Government of the Republic </ADDRESSEE>
<ACTION OF DELEGATION> is delegated to enact, </ACTION OF
DELEGATION> <TERM> within eighteen months from the date on
which this Law comes into force </TERM>, <DELEGATED/ ACT/S> one
or more legislative decrees <\DELEGATED/ ACT/S> <OBJECT
DELEGATION> laying down additional provisions of the legislation on
privacy and personal data protection, </OBJECT DELEGATION>
<GUIDING CRITERIA> complying with the following principles and
guiding criteria: a) to specify the way in which personal data used for
historical, research and statistical purposes shall be processed, taking into
account the principles found in ... </GUIDING CRITERIA>
(Law 31 December 1996, No. 676).
Krems 19 may 2004
11
CONSIGLIO NAZIONALE DELLE RICERCHE
Istituto di Teoria e Tecniche dell’Informazione Giuridica
http://www.ittig.cnr.it
Definition and structural and semantic classification
of the explicit text amending provision
object
AMENDMENT
 Part
action
 Repeal
 Substitution
 Integration
• supra-part
• article
• paragraph
• letter
• number
 Part of discourse
• sentence
• phrase
• word
Krems 19 may 2004
12
CONSIGLIO NAZIONALE DELLE RICERCHE
Istituto di Teoria e Tecniche dell’Informazione Giuridica
http://www.ittig.cnr.it
Pattern of the Sophia parser
FINE:vpred+DETX?+
RIFB+[M_ALLCAT]*+RIFE:endpos+PUNCTX?+
([E-SUCCESSIVE-MODIFIC-M]|[COME-MODIFICATO-M])?
+PUNCTX?+(INSERIMENTO:vazione|AGGIUNTA:vazione)+
((DETX+AX)|ARTICOLO_I:endnov|COMMA_I:endnov|LET
ERA_I:endnov|NUMERO_I:endnov)+(PUNCTX)?
+VIRGOLETTE:startnov+
[M_ALLCAT-PLUS-RIF]*+VIRGOLETTE:endnov
Krems 19 may 2004
13
http://www.ittig.cnr.it
CONSIGLIO NAZIONALE DELLE RICERCHE
Istituto di Teoria e Tecniche dell’Informazione Giuridica
Krems 19 may 2004
14
CONSIGLIO NAZIONALE DELLE RICERCHE
Istituto di Teoria e Tecniche dell’Informazione Giuridica
http://www.ittig.cnr.it
Applications and Future Developments of the
Project
• The tagging of the amending provision is indispensable
for the compilation of a co-ordinated text.
• The recognition of the delegation provision, is necessary
to monitor the moment in which the different
delegations attributed to the executive will expire.
• Finally, we believe that we can also apply the
methodology we have illustrated here, to projects for the
control of the quality of legislation.
Krems 19 may 2004
15
CONSIGLIO NAZIONALE DELLE RICERCHE
Istituto di Teoria e Tecniche dell’Informazione Giuridica
http://www.ittig.cnr.it
Index of the quality of the regional laws of
Tuscany
The Working Group of the Regional Council of
Tuscany has defined the concept of the quality of
laws, starting from the assumption that quality
is to be understood, as the relationship between
the text of the law and the legislative drafting
rules
Krems 19 may 2004
16
CONSIGLIO NAZIONALE DELLE RICERCHE
Istituto di Teoria e Tecniche dell’Informazione Giuridica
Experience of The Working Group of the
http://www.ittig.cnr.it
Regional Council of Tuscany
• The rules under examination were those which, having
a high technical profile, could be directly applied by
regional legislative offices.
• The analysis based on the comparison between the
application and the failure, to apply the rules within a
regional law, was conducted step by step, each one
corresponding to a qualitative aspect of the law.
• The Working Group drew attention to the rules-quality
factors which were on the whole applied with greatest
recurrence in 39 laws.
Krems 19 may 2004
17
CONSIGLIO NAZIONALE DELLE RICERCHE
Istituto di Teoria e Tecniche dell’Informazione Giuridica
http://www.ittig.cnr.it
However, in the legislation quality evaluation,
appear to be indispensable (1)
•
1.
tools for the automated recognition of natural language
so that the text structures that do not comply with the
legislative drafting rules can be identified. These tools are
even more necessary for the analysis of extended corpora,
reliable "metrics" for measuring the errors that are
found and the subsequent preparation of these
measurement in statistical indexes aimed:
at expressing the quality levels;
2.
at making comparisons.
•
Krems 19 may 2004
18
CONSIGLIO NAZIONALE DELLE RICERCHE
Istituto di Teoria e Tecniche dell’Informazione Giuridica
http://www.ittig.cnr.it
However, in the legislation quality evaluation,
appear to be indispensable (2)
• The second necessity which seems impelling is to involve
and co-ordinate centres of excellence in the fields of
documentation and legal, linguistic and statistical
processing and the control and evaluation of quality;
• The Istituto di Teoria e Tecniche dell'Informazione
Giuridica (ITTIG), the Accademia della Crusca, the Public
Law and Statistics Departments of Florence University
have decided, to collaborate together on national and
European research projects in this domain.
Krems 19 may 2004
19
Fly UP