At present, EU consular law is under legal scrutiny by the European Commission. The CARE study reveals good pragmatic application but also significant implementation problems. As a site effect of our analysis, we have developed a concept of a legal ontology for knowledge description, multilingual information retrieval and semi-automatic application of consular law using a dialogue system. First experiments show the potential of this approach.
the documents. Metadata are translated into English and French. Texts of the most relevant documents are translated into English and French as well. The database is accessible by all European citizens via the Internet (http://www.careproject.eu/database). A comprehensive report analyzes the legal framework in the EU Member States based on assessments of 27 national correspondents.
From a legal point of view, significant insufficiencies of implementation of Article 23 TFEU exist, in particular concerning legal frame work, standards of legal rules, reimbursement etc. These problems are solved in practice with a pragmatic implementation.
An ontological analysis shows that conceptualisation of consular law remains sketchy. Neither International treaties nor national laws have developed a strong terminology on consular law. Even a lexical ontology may provide important assistance.
Further, an ontology can be considered as an approach for solving the problem of multilingual (e.g. in 23 Community languages) handling of consular cases (see for the long list of functions Article 5 of the Vienna Convention on Consular Relations), taking into account the 27 different consular protection laws and policies. The ontology can provide required equivalence of concepts but can be linked also to a dialogue system. For these reasons, experimental research on legal ontologies and dialogue systems has been undertaken. The remainder of this paper is organized as follows: Section 2 describes the consular law legal information system, section 3 the ontology of EU consular law, section 4 the dynamic legal electronic commentary, section 5 first experiments and, last but not least, in section 6, tentative conclusions are presented.
.2. Legal Information System CONSUL
Handbooks in paper have long ceased to constitute best practice for
dissemination of information. Websites and information systems are able to
very nicely present the complex knowledge while coping very efficiently
with often daily updates (e.g. travel recommendations). For finding
materials, legal search constitutes an indispensable tool. Legal retrieval
remains the best solution for determining the similarity between documents
and queries
The existing CARE database already allows full text information retrieval
and browsing in the document collection. Our more powerful document
retrieval system is going to be built using Apache Lucene
Since the 1990ies, ontologies as a conceptualisation of a domain are
considered as tool for organising legal knowledge. Later, the idea of a
semantic web
In our case, we consider using two ontologies: a lexical ontology like in the
LOIS project
A thesaurus for indexing contains a list of every important term in a given domain of knowledge and a set of related terms for each of these terms. A lexical ontology builds up from this basis with works on glossaries and dictionaries, extends the relations and makes this knowledge computerusable in order to allow intelligent applications. Lexical ontologies provide this formalized description of a domain that can be understood and re-used by a knowledge system.
Based on already existing indices and sketchy conceptual structures, a lexical ontology CONSUL with about 200 legal and factual descriptors with definitions and relations has been established. Content relations will be taken from standard WordNet relations (especially hyperonymy and hyponymy). For all concepts, an ILI will be created in order to support multilingual use but also multilingual retrieval. Methodology is mostly derived from the previous LOIS project.
The Dynamic Legal Electronic Commentary (DynELCom)
The easier formalisation of knowledge and semi-automatic knowledge acquisition allow dynamic semi-automatic updating of the knowledge base. The goal is an ontological index like that in legal commentaries, however, without the textual components. It is obvious that the readability of such ontological structures is limited and will require some training. However, the exact representation of the underlying conceptual und logical structure of the legal system is much better represented.
The DynELCom CONSUL is a model of a semantic legal knowledge system. Legal knowledge is formalised with tools of the semantic web and of legal ontologies. Browsing and handling of the legal text corpus is supported by a conceptual structure with links.
The main difference to existing approaches of legal ontologies lies in the fact that world ontologies (e.g. consular factual situations) are also included in this conceptual structure. As many quite developed ontological representations of world knowledge already exist, such knowledge can be used for enrichment of an ontological representation of the legal system. A major part of the DynELCom CONSUL consists of the link structure between the facts (world ontology) and rules (legal ontology). Thus, legal reasoning is supported that may be sufficient in “easy cases and a valuable support in contradictory situations.
The formalisation of a legal knowledge domain with the DynELCom CONSUL allows also semi-automatic and automated applications. Conceptual search of links to factual and legal concepts are obvious results of this representation. This search can be supported by dialog systems that support the user in establishing relevant facts of a case. Thus, a sketchy form of automated legal reasoning can be offered, e.g. a “simplified legal syllogism”. The facts of a case are properly refined by a dialog system leading to a factual concept but also a legal concept.
The DynELCom CONSUL faces the dynamics of the legal system. The indispensable indexing and analysis process is supported by semi-automatic categorisation and text analysis. Computational linguistics, text extraction, document categorization and text summarization tools are now sufficiently powerful so that good results can be achieved in very short time. The analysis of the DynElCom is based on a co-operative work model between the man and the computer. The legal information system provides the basis for the commentary. The knowledge base with the ontology and semi-automatic text analysis provides extensive knowledge of the text corpus of the legal information system. Software tools are information retrieval, hypertext, knowledge management, text summarisation, text categorisation and the inference machine. Manually, ontologies have to be established and maintained, semi-automatic indexing must be constantly fine-tuned and inference engines must be supervised. Such work is presently done by legal authors and practitioners. With the DynELCom CONSUL, a concentration of such analysis takes place in a semi-automatic way. The main advantage is real time delivery, higher quality and lower costs.
The main advantage of the DynELCom CONSUL seems to be obvious: in “easy cases”, much of the work can be automated. Consular services would become cheaper with higher quality. Existing pressures on public budgets may lead to cuts in consular networks. With semi-automatic systems, much work can be outsourced to other consular posts of other Member States or honorary consuls.
Text corpus: The basis for the text corpus is the CARE project database. Only few modifications are envisaged; mostly hypertext links to the ontology, visual representations and a list of document types. Ontology CONSUL: The ontology consists of a legal ontology, a world ontology and links (anchors) between the legal and the world ontology. Elements of the ontology are 3 types of frames: legal frame, fact frame and anchor frame. A frame contains a header, definitions (with sources), classification codes, and relations (to other frames, e.g. synonym, homonym, polysem, hyponym, hyperonym, antonym etc. but also to an anchor). The anchor frame can best be described as a citation with a header, the identification (abbreviation or number) and links to facts and legal concepts. For the representation, existing standards of the semantic web and legal ontologies are implemented; in particular OWL, RDF and LKIF. This first step with a frame-like representation of legal concepts, factual concepts and the anchors between facts and rules will be followed by a second step that intends a more sophisticated ontological representation of the legal system. This representation focuses on space, persons, actions, material rules and procedural rules.
Action space for persons will be the real space and the cyberspace. Persons can be natural persons (and quasi-persons, e.g. robots or software agents). Objects in the space are physical objects (thinks), energy and quasiphysical objects (e.g. web store). Actions can be physical processes (actions or non-actions in real space or quasi-physical processes (actions on the web). Mental objects and mental processes consist of combinations of these elements. Due to social practice, such “virtual” sets are considered as a unified object or process (e.g. organizations, enterprises, associations, families etc.) Law builds on the existing physical and social structure of persons, objects and processes but modifies it or adds particular elements. Persons can be natural or legal (e.g. limited company, international organization, state), objects are physical, mental or legal, and actions are physical, mental, or legal. It is obvious that the differences between the social reality and law (as representation as the world should be) are a high interest in any legal system. The representation is structured in concepts (an ontology), rules and factual situations. Isomorphism is respected via direct links to norms but also its logical representation. Legally relevant links between the world ontology and the legal ontology provide support for legal reasoning (e.g. possible factual situations or legal consequences of certain facts). In the LKIF terminology, such a function is called anchors (LRI-Core Ontology). They provide an anchor function as links between the social spectrum of actions and legally.
Knowledge acquisition tools: Text extraction and summarisation tools are
decisive for the knowledge acquisition. The tools consist of a knowledge
base containing the extraction, summarization and classification rules with
header, rule, definition and relations and several tools for semi-automatic
text analysis providing information on relevant documents, extract
important text passages, classify documents, deliver definitions etc.
We have developed prototypes and applications on corpora-based text
analysis for about 20 years now. Due to space restrictions, we can provide
only a very short overview of the methods. A pre-defined list of descriptors
can be checked against a text corpus with the KONTERM method
Dialogue system: Such a system is intended to converse with a human in a
coherent structure
Due to time and financial restrictions, the implementation of the DynELCom CONSUL has mostly remained a concept. However, due to our ongoing involvement in the CARE project, we have worked for more about one year on a partial experimental application. The following presentation provides first experiments.
Existing text corpora (RIS, EUR-Lex, CARE) and ontologies resulting from our daily work with European, international and Austrian law forms the basis for these experiments. For Austrian and European law, we have established an ontology with a sufficient granularity of an ontological representation of a jurisdiction: about 10,000 thesaurus entries, 5,000 citations, up to 200 document types, a classification structure, 100 text extraction and summarization rules. This meta data is stored and updated in a database with different types of knowledge frames: Fact and legal descriptors: header, definition (with sources), examples (with sources), relations (synonym, homonym, polysem, hyponym, hyperonym, antonym etc.), classification, other information.
Anchors: header, identification (abbreviation or number), synonyms, classification, author, links, other information.
Document types: header, identification (abbreviation), use, format, other information.
Classification: header, code, definition, relations, other information. Extraction and summarization rules: header, rule, definition, relations, other information.
Concepts: header, definition (with sources), related thesaurus entries and citations, relations (synonym, homonym, polysem, hyponym, hyperonym, antonym etc.), classification, legal conceptual structure (ontological model), other information.
Rules: header, quasi-logical expression, source, type, classification, legal conceptual structure (ontological model), other information. Procedures: header, flowchart, source, type, classification, legal conceptual structure (ontological model), other information.
This ontology was extended to the consular and diplomatic protection. The following examples may show the lexical ontology (note: an (L), (F) or (A) is added to the header in order to distinguish between legal and fact descriptors as well as anchors). The attribute “legal conceptual structure” indicates relevant branches of law.
Header: Evacuation (L) Definition: In case of a catastrophe (e.g. earthquake), EU Member States will evacuate their citizens (and family members) as a matter of law or policy. EU Member States will co-operate and support each other for this goal (Art. 23 TFEU). Source: Article 23 TFEU, national laws, CARE project report Relations: BT catastrophes (F), BT consular assistance (L), catastrophes (A) Classification: CAT:EVA Legal conceptual structure: consular assistance, catastrophes Other information: none
Header: 2010 Earthquake in Haiti (F) Definition: Earthquake of 12 January 2010 with an epicentre near the town of Léogâne affecting about 3 million people in Haiti.
Relations: Catastrophe (F), evacuation (L) Source: English Wikipedia Classification: CAT.EAR Legal conceptual structure: Evacuation (L) Other information: none
Header: Catastrophes (A) Links: Terrorism (F), earthquake (F), tsunami (F), hurricane (F), flooding (F), international conflict (F), consular assistance (L), Article 23 TFEU (L), evacuation (L) etc. At present, we are in the process of finishing the first prototype of this representation of Austrian consular law. The next step would be a verification of the conceptual structure using the knowledge acquisition and text analysis tools. Such a process is very time-consuming and requires financial resources not available so far. However, the existing ontology provides already a very helpful tool for legal work as it represents legal and fact concepts and its links.
In this paper, we have given an outline of a system for semi-automatic application of consular law in a multilingual and multinational environment, focusing on the underlying legal ontology. For the moment, we are working on a more sophisticated and extended ontological representation.
The CARE Project is funded by the European Commission, Grant No. JLS/2007/FRC-1/50-30-CE-0226854/00-31.