The Legal Knowledge Interchange Format (LKIF), is a Semantic Web based language for representing legal knowledge in order to support modeling of legal domains and to facilitate interchange between legal knowledge-based systems.

LKIF has been designed as a knowledge interchange format targeted particularly for applications in legal domains. Its architecture is a resulting compromise between a range of objectives and requirements, which in many cases could not have been satisfied simultaneously. The dominant trade-off underlying the design process concerned conflicting demands for an increased expressiveness of the language on the one hand, and its tractability and compatibility with established standards on the other. In this section we present a brief overview of the structure of LKIF and specify its fragments, which will be addressed in the following sections. In general, LKIF provides a direct support for representing three types of knowledge, which have been identified as most indispensable to law and legal reasoning: terminological knowledge, legal rules and normative statements.

Following the adopted requirements, LKIF complies with the Semantic Web and Knowledge Representation philosophy of a layered representational structure. Predominantly, we can distinguish between terminological and rule layer of the language, which constitute its two expressive dimensions. Furthermore, LKIF vocabulary reserves a special set of deontic concepts aimed for modeling normative statements.

Terminological Knowledge: The layer of terminological knowledge is supported in LKIF through the Web Ontology Language (OWL), one of the recommended XML-based standards for Semantic Web. OWL allows for explicit and formal representation of meaning of terms and relations between them, thus enabling construction of machine-accessible models of ontologies underlying particular domains of knowledge. Ontologies specified in OWL can be stored as OWL files. The terminological support provided by LKIF is extended with the LKIF-Core ontology of basic legal concepts, designed as another part of the work package.

Legal Rules: Whereas terminological layer has easily fit into the recommended Semantic Web standards, representation of legal rules has required more sophisticated formalism. Initially, one serious candidate language has been considered: SWRL, put forward as a proposal for Semantic Web rules-language, combining sublanguages of the OWL Web Ontology Language (OWL DL and Lite) with those of the Rule Markup Language (Unary/Binary Datalog)
SWRL has turned out insufficiently expressive to capture some of the essential peculiarities of legal reasoning. For that reason, a partly novel rule formalism, called LKIF rules, has been developed and incorporated in LKIF.

LKIF rules extend SWRL with support for negation and defeasible reasoning. The rule layer provides a language expressive enough to model legal rules in a way which comes much closer to the ideal of isomorphic modeling, i.e. in a way which reflects the structure of the rules in a legislation. The rules layer supports rules with exceptions, assumptions, and exclusionary conditions, and enables meta-level information about rules to be represented, such as date of enactment, which is used in other rules, such as lex posterior, to reason about rule priorities.

The argumentation module in LKIF provides a conceptual framework, underpinned by an argumentation-theoretic semantics, for representing rule constructs as OWL objects. Rules can be hence embedded in OWL files alongside terminologies. It has to be noted that after completion of D1.1, W3C announced a proposal for the Rule Interchange Format, a new Semantic Web standard for combining rules with ontologies. This formalism, though not originally considered, should be taken into account and assessed from the perspective of a possible reorientation of the LKIF rule layer in the future.
Normative Statements: Normative statements are given a direct support via the Norm module included in LKIF-Core ontology. No particular deontic logic is imposed on LKIF representation. Such an approach is justified by both theoretical and practical reasons. On the one hand there is a little convergence in the field of deontic logic to the standard system of reasoning, on the other — industrial applications do not typically require a reference to any concrete deontic logic. Instead, the module contains minimally restricted definitions of deontic concepts and properties, based on the notion of subjunctive betterness, which are intended as basic constructs for modeling various deontic settings. Nevertheless, there exists a preferable mapping of normative statements into OWL representation, which preserves some essential properties of the modal frames semantics, associated with deontic logics.

To summarize, the Legal Knowledge Interchange Format is an OWL ontology of legal concepts allowing legal knowledge bases, encompassing specific terminologies, LKIF rules, and normative statements, to be represented in OWL and stored as OWL files. Let us finally introduce the notion of an LKIF reasoner, an LKIF file and of using LKIF:

• Definition 1 (Using LKIF) To use LKIF is to refer to an entity in the LKIF namespace.
• Definition 2 (LKIF File) An LKIF file is a file which uses one of the concrete syntaxes described in this section, is limited to one of the expressive fragments identified in this section, and refers to an entity in the LKIF namespace.
• Definition 3 (LKIF Reasoner) An LKIF reasoner is a reasoner that interprets a set of RDF triples that refer to an LKIF entity, and that is limited to one or more of the identified expressive fragments, in accordance with LKIF semantics.

The interested reader is referred to Deliverable 1.1 “Specification of the Legal Knowledge Interchange Format” and Deliverable 4.1 “The Legal Knowledge Interchange Format” (a revision) for an in-depth description of the LKIF language, or the “Summary of D1.1″ .