RAWE: A Web Editor for
Rule Markup in LegalRuleML
Monica Palmirani1, Luca Cervone1 Octavian Bujor1, Marco Chiappetta1
1
CIRSFID, University of Bologna.
{monica.palmirani, luca.cervone, octavian.bujor, marco.chiappetta}@unibo.it
Abstract. This paper presents a Web editor (RAWE: Rules Advanced Web
Editor) for marking up legal rules starting from legally binding texts. The Web
editor exploits the legal information embedded in the Akoma Ntoso markup, in
combination with and XML techniques, so as to help the legal-knowledge
engineer model legal rules and convert them into LegalRuleML, an OASIS
XML standard candidate.
Keywords: Legal Reasoning, Akoma Ntoso, LKIF-core, LegalRuleML.
1. Introduction
This paper presents a Web editor for marking up legal texts in a legal document’s
main structure, normative references, and legal metadata using the Akoma Ntoso [2]
[13] [25] XML standard, now undergoing the OASIS standardization process. The
same Web editor exploits the legal information embedded in legal markup, in
combination with XML techniques, to help the legal-knowledge engineer model legal
rules using a logic formalism and convert them into LegalRuleML [1] [23][24],
another OASIS XML standard candidate. The two standards—Akoma Ntoso and
LegalRuleML—are complementary in implementing the legal-knowledge modelling
and representation of legal documents. The main goal of the RAWE Web editor is to
provide a tool capable of managing in an integrated way the advantages of the Akoma
Ntoso and of LegalRuleML, applying the isomorphism principle [3][9][22] to
connect, as far as possible, legally binding textual provisions with the logic formalism
expressed using rules. Usually, AI&Law experts are too focused on the task of
applying a logic formalism to achieve isomorphism, but the legal experts (judges,
lawyers, and administrators) are interested in verifying the results of the legal
reasoning engine and in finding evidence in the legally binding text.
Secondly, a legal text changes over time, and so the rules need to be updated
accordingly. If the isomorphism principle is not applied, it is quite difficult to
determine whether those rules need to be updated. The RAWE editor helps to
maintain text and rules aligned and to minimize manual markup activity.
Thirdly, the aim of the RAWE is to show how it is possible to export
LegalRuleML in RDF serialization to favour Linked Open Data interoperability.
Finally, in the future, the same editor will export LegalRuleML files in other
proprietary languages, like SPINdle [15] or Carneades [8][11], so as to permit legal
reasoning.
2. From Open Text to Open Rules
The first point to be made in clarifying the goals RAWE would like to achieve is to
draw a distinction among three conceptual layers: norms (abstract mandatory
commands concerning rights or duties), textual provisions (sequences of texts), and
rules (rendering of the text into logical rules).
A norm, following Kelsen’s definition [14], is an abstract mandatory command
concerning rights or duties. A norm is usually expressed in writing using legal texts or
in an oral way (e.g., a social norm, an oral contract) or in other representations (e.g.,
symbolic road signs).
Textual provisions (or simply provisions) are the instantiation of general norms
in one possible textual representation (a sentence, article, or paragraph).
Legal rules are interpretations of one or more provisions formalized using logical
rules in the form of antecedent and consequent. Sometimes several provisions will
form a single rule, or a single provision may include multiple rules.
Usually, in the state of the art, AI&Law scholars focus their attention only on the
rule modelling and on the foundational logical theory, and apart from the
isomorphism principle [3], the connection with the text over time and the ontology
aspects have been neglected. There is an important theoretical debate in the AI&Law
community on the interpretation of the legal textual provisions expressed in natural
language and on the canonization of rules using logical formalisms [4]. The prevalent
theory is now oriented towards hybrid interpretation [27] (rather than pure textualism,
or pure interpretation). We want to make visible in the text the “evidence” that there
is a minimal but reasonable interconnection, following the legal theory of
interpretation, with a logical rule in a formal representation. This exercise sometimes
forces the legal-knowledge expert to split the original provision into two or more
rules, or to duplicate the rules, or to compress several sentences into a single rule. In
this scenario, we have to manage an N:M relationship among norms, textual
provisions, and the ontology that we want to capture and represent maintaining a
strong separation among these three levels.
Nevertheless, it is obvious that the isomorphism approach alone presents some
exceptions and limitations that need to be balanced in a reasonable way. We have at
least three cases where the legal rules have no textual link: (i) when we have implicit
rules deriving from the general principles of the legal system (e.g., lex superior, lex
specialis, lex posterior); (ii) when the legal-knowledge engineer includes a personal
interpretation as a summary of his/her expertise; and (iii) when the legal-reasoning
engine produces rules. In these cases the Web editor provides metadata to distinguish
those rules deriving from the legal text from those that are a free interpretation of the
rules’ author. The RAWE editor permits multiple interpretations of the same legal text
and makes it possible to follow the isomorphism principle, but also to derogate from it
if need be.
Finally, the Web editor exports all the metadata in RDF format to favour the
interconnection of Legal Open Data with Linked Open Data. The goal is to release
RDF triples about the rule knowledge base, in such a way as to connect that with
other datasets available in the Linked Open Data Cloud. This permits more-effective
filters of the legal resources in the Semantic Web domain (e.g., geo-localizing legal
resources on the map using the jurisdiction and the temporal metadata filter to find the
legal rules relevant to a given context, such as environment law or construction law).
Fig. 1 – Scenario of relationships among different layers
in legal knowledge modelling
3. Related Work
The AI&Law community itself [27] has spent the last two decades modelling legal
norms using different logics and formalisms, usually fed manually to a legal-
reasoning engine. Some visual tools [11] or editors [15][19] in the past have been
developed to model rules, but the methodology used starts from a reinterpretation of
the legal source text by a legal-knowledge engineer who extracts the norms, applies
models and theory using a logic representation, and finally represents them with a
particular formalism. The RAWE approach is different: it starts from the legal text
marked up in some Legal XML standard and, exploiting the text’s regularity, detects
some metadata that is also useful for modelling rules.
Over the last decade, several Legal XML standards have arisen for describing
legal texts (Akoma Ntoso, CEN Metalex [5] [16]) and rules (RuleML, RIF, SWRL,
etc.), but the two communities are mostly separated, and they pursue their goal
separately. In the meantime, the Semantic Web, and in particular legal ontology
research, combined with the NLP extraction of semantics, has given a great impulse
to the modelling of legal concepts [17][18][20][7][6][12][26]. In this scenario there is
an urgent need to close the gap between the text description, represented using XML
techniques, and the norms formalized with logical rules, this in order to realize an
integrated and self-contained representation. There are three main reasons:
• Legal knowledge is currently presented in a disjointed way in the original text that
inspired the logical modelling. This disconnection between legal-document
management and the logical representation of the embedded rules strongly affects
the real usage of the legal-document knowledge in favour of citizens, public
administrations, and businesses (e.g., contracts, insurance regulation, banking soft
law).
• Management of changes undergone over time by legal documents—especially acts,
regulations, and contracts—that by nature are variable and subject to frequent
modifications, significantly affecting the coordination between the text and the
rules that should be remodelled.
• The legal validity of the text as authentically approved by the competent entities
(e.g., contractors) should be preserved across all manipulations. On the other hand,
it is important to connect legal document resources, which themselves include
many legality values (e.g., authenticity, integrity, evidence in trial, written form,
etc.), with the multiple interpretations coming from legal-knowledge modelling.
Certainly, one of the main challenges over the last five years has been to acquire
the ability to capture, with the help of NLP techniques, all the relevant legal
knowledge embedded in a legal document and to represent it in an appropriate formal
model. However, there hasn’t been significant progress on the state of the art in this
respect, especially in languages other than English. So it is important to improve the
user interface technique to help the legal-knowledge expert to easily model legal rules
and prepare an environment for a future NLP integration. RAWE is the only Web
editor in the state of the art that can model legal texts and rules in a coordinated and
consistent way using a WYSIWYG interface exploiting two important legal XML
standards: Akoma Ntoso and LegalRuleML.
4. Akoma Ntoso and LegalRuleML Synergy
As mentioned before, Akoma Ntoso and LegalRuleML are two XML standards for
modelling and representing legal documents. RAWE can coordinate the knowledge
captured with these two standards so as to help the end user mark up the legal rules
using a logic formalism enriched with temporal parameters.
Akoma Ntoso is specifically designed to model a legal document’s structure and
legal metadata, like the preface, preamble, sections, conclusions, normative
references, dates, and signatures. The Akoma Ntoso metadata block additionally
defines the conditions under which the legal textual fragment is valid, effective and in
force, while also defining jurisdiction, the document’s authority, and other relevant
legal metadata, like modifications. All those metadata are also significant in defining
the context of a legal rule, helping the legal reasoning engine filter the rules pertinent
to a particular case (e.g., infringement of the rule at a given date in 1999).
The following example displays as the block that
defines the interval of efficacy and enforceability of Section 504 of the US Code.
... meta data about the legal document ...
Cover page content
... the preface of the document ...
... the preamble of the document ...
... the normative part of the document ...
LegalRuleML is designed to model in logical formalism the norms expressed in a
legal text. It does so especially using deontic operators: obligation, right, permission,
prohibition. LegalRuleML is also intended to define the context for each rule by
providing a set of metadata like the temporal parameters, the original textual sources,
the jurisdiction, the author, and the authority of the rules. The fragment below shows
the main structure of a LegalRuleML document composed of different metadata
blocks defining the author who modelled the text into rules (),
recording the original legal resources IRI (), and providing the
temporal parameters (), the context, and each
rule’s date of creation (). The provides the environment in which the rules are
valid (time, author, jurisdiction, etc.).
general
definition of
m.palmirani
definition of
definition of
instant time
1999-12-
09T00:00:00.0Z
definition of
intervals and
...
... rule base block
...
Entering all the information manually for each rule is a really
time-consuming task, especially when the legal text has gone through several
modifications over time. Moreover, it is difficult to maintain consistency between
legal textual provisions and rules in the dynamicity of the legal system. For this
reason the RAWE Web editor exploits the information embedded in the Akoma Ntoso
text proposition (e.g., section, article), and it reuses those data to define the context of
the rules when accurately connected to the legal provision.
The following example presents a fragment of Section 504 of the US Code
concerning copyright infringement and the related rules. Section 504 is presented in
the version updated at time t5, which in Akoma Ntoso is defined in the
block.
When the end-user selects a portion of the legal text with the mouse in the Web
editor window, all the related metadata recorded in Akoma Ntoso are detected and
exported in LegalRuleML to model the rules.
The following example shows in the two standards (i) the correspondence among
temporal events; (ii) the correspondence among temporal intervals; and (iii) how it is
possible to reuse the Akoma Ntoso information in LegalRuleML (compact form).
i) Event definition in the Akoma Event definition in LegalRuleML,
Ntoso metadata block automatically extracted from the
Akoma Ntoso text using mouse-over
type="amendment" date="1999-12-
Intervals definition in Akoma Intervals definition in LegalRuleML
Ntoso in the metadata block
source="#palmirani"> key="e2-e">
keyref="#t6"/>
Context of rule1 in LegalRuleML,
automatically built using Akoma
Ntoso information
Text in Akoma Ntoso Rule definition in LegalRuleML
connected to the textual provision
selected by mouse-over
clsc">
Statutory
Damages.
clsc-lst1"> min Pay
sec504-clsc-lst1-pnt1">
(1)
X
-Except as provided by clause 750
(2) of this subsection, the
copyright owner may elect, at any
time before final judgment is
actual damages and profits, an Pay max
all infringements involved in the
action, with respect to any one
work, for which any one infringer X
is liable individually, or for 30,000
which any two or more infringers
are liable jointly and severally,
in a sum of not less than
period="#t5">$750 or more
than
period="#t5">$30,000 as
the court considers just. For the
purposes of this subsection, all
the parts of a compilation or
derivative work constitute one
work.
5. From LegalRuleML Meta-model to RDF Serialization
LegalRuleML was designed based on a meta-model1 that defines relationships among
different classes of the elements in the XML-schema. For helping this approach the
technical author of the XML-schema (Tara Athan) implemented also several rdfs
schemas. The following fragment of rdfs schema shows the relationship among the
element and the property . Following this
approach all the elements that start with lower case are edges and the elements that
start with upper case are nodes of a graph.
Role
The class of roles played by agents relative to
LegalRuleML things.
appliesRole
1 Meta model is now under revision and the authors take this version from the OASIS
repository: https://tools.oasis-open.org/version-
control/browse/wsvn/legalruleml/trunk/schemas/?rev=71&sc=1
A role applied to the targets by
the subject association or rule context.
Using this meta-model it is possible to extract some relationships among elements.
Some assertions in RDF format about the knowledge base rules are possible
especially from the . These assertions build a set of RDF triples
useful for improving information retrieval of the legal rules, and related legal textual
sources, in the Semantic Web. The contextualization of the legal rules (e.g.
Jurisdiction, Author, Authority, etc.) permits to create enriched connection with the
Linked Open Data Cloud (e.g. geo-localization of the legal rules on the maps):
The same mechanism should be applied to the other assertions included in the
.
6. RAWE Functionality
RAWE permits the following functionalities:
• Authentication of the end-user and customization of the environment according
with the personal profile (e.g., legal system, legal tradition, legal guidelines);
• Multilanguage interface and environment;
• Customized interface and buttons on the basis of the user profile;
• Mark-up of a legal text with Akoma Ntoso standard using parsers to automatically
detect the normative references, dates, metadata, and structure of legal documents;
• Record of the XML files in the eXist repository [21];
• Tree of the marked-up elements;
• On-the-fly view in Akoma Ntoso and in LegalRuleML;
• Conversion and export in PDF, XML, ePub, or RDF format;
• Web editor environment with WYSIWIG interface;
• Undo function;
• Contextual functionalities based on the XML tree and XML-schemas;
• Mouse-over for detecting the metadata of a portion of legal text and reuse for
modelling legal rules;
• Toolbar for marking up the document’s structure;
• Toolbar for marking up legal rules.
Fig. 2 – RAWE Web editor for marking up legal texts and normative rules
There are some critical points that we have faced in the RAWE implementation
using HCI techniques:
• Contextual Composition of the Rule. In LegalRuleML we have three groups of
rules: Prescriptive, Constitutive and Behaviors. Each group permits some particular
modeling following the legal theory (e.g. Prescriptive rule is a sequence of deontic
operators, Penalty needs a separate regime, Constitutive rule doesn’t include
deontic operators, etc.). For this reason RAWE needs to take in consideration the
LegalRuleML prescriptive grammar constraints and lead the end user to compose
the rules correctly.
• Reparation is a binary relationship between a penalty and a prescriptive rule or
violation. So we found a smart interface way to select the two parts of the
relationship and to connect them to each other.
• Metadata in Context. If we need to refine or readjust the context and the related
metadata, we need a new toolbar and panel. RAWE permits to readjust the
metadata imported by Akoma Ntoso and to add new ones.
• Extra isomorphism rules. Sometimes we need to include extra rules not directly
linked to the legal text. RAWE permits to model this particular situation.
However other some critical issues need to be addressed in the future:
• Ontology. Some elements of the rule modeling need to be enriched with the
definitions of an external vocabulary or ontology (e.g. LKIF[10]).
• Key. We need to create a naming convention to harmonize the ID definition.
• Meta-Rules. In the future LegalRuleML will be also be able to manage meta-rules
(rules about other rules), and we need to find a mechanism for linking rules as
antecedents and consequents.
• Multiple interpretation. In this version of the editor is not possible to have
multiple interpretations of the same legal textual document fragment.
• Granularity. For now the granularity of the isomorphism is on the rule. In the
future we will be able to also manage the same functionality on the body, head, and
atom.
Fig. 3 – RAWE conversion of a rule in the LegalRuleML standard
7. The RAWE Architecture
RAWE is a specialized Web editor developed using several open-source technologies,
such as Sencha ExtJS 4.1 and TinyMCE.
Sencha ExtJS is an MVC framework that makes it possible to build an
extraordinarily rich Web application. It supplies the instruments with which to easily
develop the core of the application based on the Model View Controller pattern, and,
moreover, it comes with a big range of user interface widgets. The other core strength
of ExtJS lies in its component design. If the developer needs a new component that is
not yet developed, the default components can be extended and the result is
encapsulated in the default components. ExtJS is also completely cross-browser, so it
is possible to deliver the application on a wide range of browser and operating
systems. The latest smart phone and tablet browser are also covered, so it is possible
to use ExtJS-based applications with touch screens and gestures.
TinyMCE is a platform-independent Web-based Javascript HTML WYSIWYG
editor control. It can convert HTML text area fields or other HTML elements into
editor instances. We integrated it into ExtJS, developing a new component for the
framework. The component retains all the functionality of the TinyMCE editor, but
the effects of those functionalities are intercepted by the core of the ExtJS application.
With this strategy, each event handled by the editor simply fires other events handled
by the other components of the application. This means that there is no specific
semantic on which TinyMCE itself
relies, and TinyMCE can be
substituted on demand with other
open-source WYSIWYGs.
The editor uses the HTML5
standard in order to mark up
documents. When an element is
marked up, it is wrapped by a
generic HMTL element (such as
span or div), and various classes are
assigned to it in order to give to it
semantic meaning for the editor
itself and for the tool in charge of
translating it into the desired
document format. This means that
there is not a meta markup language
in the middle of a translation from
HTML to another document format,
and this carries the benefit of
preventing data loss and having
immediate access to the HTML version of the document without further conversions.
8. Conclusion
We have presented RAWE, a Web editor for marking up legal rules exploiting the
previous markup of legal texts in Akoma Ntoso. RAWE is developed to enable
application of the isomorphism principle; nevertheless, it is also open to the addition
of rules not properly linked with the legal textual provisions, this in order to permit
multiple interpretations or the inclusion of implicit rules. RAWE transforms all the
rules in LegalRuleML and it saves them in a native XML repository, eXist. It is also
possible to export the outcomes to a XML file. Finally, RAWE can convert in RDF
the for creating a repository capable, in the future, of implementing
an endpoint SPARQL for managing a better filter of legal resources in the Linked
Open Data Cloud. Future work will be focused on the critical points stressed in the
paper for managing advanced features.
References
[1] Athan T., Boley H., Governatori G., Palmirani M., Paschke A., Wyner A.: OASIS LegalRuleML. In
Bart Verheij, ed, Proceedings of 14th International Conference on Artificial Intelligence and Law
(ICAIL 2013). ACM, 2013.
[2] Barabucci G., Cervone L., Palmirani M., Peroni S., Vitali F.: Multi-layer Markup and Ontological
Structures in Akoma Ntoso. In: LNCS 6237/2010, pp. 133-149, Springer, 2010.
[3] Bench-Capon T. and Coenen F.: Isomorphism and legal knowledge based systems. Artificial
Intelligence and Law, 1(1):65–86, 1992.
[4] Boella G., Governatori G., Rotolo A., Torre L.V.D.: A Formal Study on Legal Compliance and
Interpretation. ;In AICOL Workshops(2009), Springer, 162-183, 2011.
[5] Boer A., Hoekstra R., de Maat E., Hupkes E., Vitali F., Palmirani M., Rátai B.: CEN Metalex
Workshop Agreement (2009-08-28 proposal). http://www.metalex.eu/WA/proposal.
[6] Breuker J., Boer A., Hoekstra R., Van Den Berg C.: Developing Content for LKIF: Ontologies and
Framework for Legal Reasoning, in Legal Knowledge and Information Systems, JURIX 2006, pp.41-
50, ISO Press, Amsterdam, 2006.
[7] Brighi R., Lesmo L., Mazzei A., Palmirani M., Radicioni D.: Towards Semantic Interpretation of
Legal Modifications through Deep Syntactic Analysis. JURIX 2008: 202-206, 2008.
[8] Ceci, M., and Gordon, T. F.: Browsing case-law: An application of the carneades argumentation
system. In Proceedings of the RuleML2012@ECAI Challenge, H. Aït-Kaci, Y.-J. Hu, G. J. Nalepa,
M. Palmirani, and D. Roman, Eds., vol. 874, pp. 79-95.
[9] Gordon T. F., Governatori G., Rotolo A.: Rules and Norms: Requirements for Rule Interchange
Languages in the Legal Domain. RuleML 2009: pp. 282-296, Springer, 2009.
[10] Gordon T. F.: Constructing Legal Arguments with Rules in the Legal Knowledge Interchange Format
(LKIF). In: Computable Models of the Law, Languages, Dialogues, Games, Ontologies (2008), pp.
162-184, Springer, 2008.
[11] Gordon, T. F.: The Carneades web service. In Computational Models of Argument – Proceedings of
COMMA 2012, B. Verheij, S. Szeider, and S. Woltran, Eds., IOS Press, pp. 517-518.
[12] Hoekstra R., Breuker J., Di Bello M., Boer A.: The LKIF Core Ontology of Basic Legal Concepts. In:
Casanovas P., Biasiotti M.A., Francesconi E., Sagri M.T. (eds.), Proceedings of LOAIT 2007, 2007.
[13] http://www.akomantoso.org/ naming convention of the URI
[14] Kelsen H.: Reine Rechtslehre, 2d. ed., Wien, 1960.
[15] Lam H., Governatori G.: The Making of SPINdle. RuleML 2009 proceeding, pp. 315-322, 2009.
[16] Lupo C., Vitali F., Francesconi E., Palmirani M., Winkels R., de Maat E., Boer A., and Mascellani P:
General xml format(s) for legal sources - Estrella European Project IST-2004-027655. Deliverable
3.1, Faculty of Law, University of Amsterdam, Amsterdam, The Netherlands, 2007.
[17] Mazzei A., Radicioni D., Brighi R.: NLP-based extraction of modificatory provisions semantics.
ICAIL 2009: pp. 50-57, ACM, 2009.
[18] Mommers L.: Ontologies in the Legal Domain. In: Poli R., Seibt J. (eds.), Theory and Applications of
Ontology: Philosophical Perspectives, Springer 2010, pp. 265-276, 2010.
[19] Palmirani M., Brighi R.: An XML Editor for Legal Information Management. Proceeding of the
DEXA 2003, Workshop on E-Government, Praga, 1-5 September, pp. 421-429. Springer-Verlag
Berlin Heidelberg, 2003.
[20] Palmirani M., Brighi R.: Model Regularity of Legal Language in Active Modifications. AICOL
Workshops 2009: pp. 54-73, Springer, 2009.
[21] Palmirani M., Cervone L.: Legal Change Management with a Native XML Repository. A cura di G.
Governatori. Legal Knowledge and Information Systems. JURIX 2009. The Twenty-Second Annual
Conference. Rotterdam. 16th-18th December 2009,pp. 146-156, Amsterdam: ISO press, 2009.
[22] Palmirani M., Contissa G., Rubino R: Fill the Gap in the Legal Knowledge Modelling. In Proceedings
of RuleML 2009, pp. 305-314, Springer, 2009.
[23] Palmirani M., Governatori G., Rotolo A., Tabet S., Boley H., Paschke A.: LegalRuleML: XML-Based
Rules and Norms. RuleML America 2011: 298-312, Springer, 2011.
[24] Palmirani M., Governatori G., Rotolo A., Tabet S., Boley H., Paschke A.: Legal-RuleML: XML-
Based Rules and Norms. RuleML America, Springer, 2011, pp.298-312.
[25] Palmirani M.: Legislative Change Management with Akoma-Ntoso, in Legislative XML for the
Semantic Web, Springer, Law, Governance and Technology Series Volume 4, 2011, pp 101-130.
[26] Sartor G.: Legal Concepts as Inferential Nodes and Ontological Categories. In Artif. Intell. Law 17(3)
2009, pp. 217-251, 2009.
[27] Sartor G.: Legal Reasoning: A Cognitive Approach to the Law. Vol. 5. Treatise on Legal Philosophy
and General Jurisprudence. Berlin: Springer, 2005.
[28] Vitali F., Palmirani M.: Akoma Ntoso Release Notes. [http://www.akomantoso.org]. Accessed 5 July
2013.