Bridging the gap between Legal Practitioners and Knowledge Engineers using semi-formal KR
The use of Structured English as a computation independent knowledge representation format for non-technical users in business rules representation has been proposed in OMGs Semantics and Business Vocabulary Representation (SBVR). In the legal domain we face a similar problem. Formal representation languages, such as OASIS LegalRuleML and legal ontologies (LKIF, legal OWL2 ontologies etc.) support the technical knowledge engineer and the automated reasoning. But, they can be hardly used directly by the legal domain experts who do not have a computer science background. In this paper we adapt the SBVR Structured English approach for the legal domain and implement a proof-of-concept, called KR4IPLaw, which enables legal domain experts to represent their knowledge in Structured English in a computational independent and hence, for them, more usable way. The benefit of this approach is that the underlying pre-defined semantics of the Structured English approach makes transformations into formal languages such as OASIS LegalRuleML and OWL2 ontologies possible. We exemplify our approach in the domain of patent law.
💡 Research Summary
The paper tackles a long‑standing problem in legal informatics: the gap between domain experts (lawyers, patent examiners, judges) who are comfortable expressing rules in natural language, and knowledge engineers who need formal, machine‑readable representations for automated reasoning. While standards such as OASIS LegalRuleML, the Legal Knowledge Interchange Format (LKIF), and OWL‑based legal ontologies provide the necessary logical rigor, they are far too technical for non‑programmers. To bridge this divide, the authors adapt the Structured English (SE) approach originally defined in the Object Management Group’s Semantics and Business Vocabulary Representation (SBVR) specification. SE is a “computational‑independent” language that retains the readability of ordinary English while constraining vocabulary and syntax to a pre‑defined set of terms, fact types, and modal operators. This ensures that the meaning of each statement is unambiguous, yet the author does not need to learn a formal logic language.
The authors present KR4IPLaw (Knowledge Representation for Intellectual Property Law), a proof‑of‑concept system that enables legal practitioners to author rules in SE and automatically translate them into two widely‑used formal languages: LegalRuleML (the XML‑based rule interchange format of OASIS) and OWL‑2 ontologies. KR4IPLaw’s architecture consists of three layers. The first layer is a web‑based authoring interface that offers auto‑completion for a curated legal‑term dictionary and a set of allowed relational predicates (e.g., “is‑a”, “has‑property”). This prevents the introduction of undefined terminology. The second layer parses the SE input with an ANTLR‑generated grammar, producing an internal meta‑model that mirrors SBVR’s concepts of Vocabulary, Fact Type, and Business Rule. Each node in the meta‑model is annotated with semantic tags that capture modality (obligation, permission, prohibition) and logical structure (condition‑consequence). The third layer is a transformation engine that maps the meta‑model to LegalRuleML and OWL‑2 constructs. For instance, SE modal verbs such as “must” or “shall” become
To evaluate the approach, the authors selected a concrete use‑case from patent law: the assessment of novelty (newness) of a claimed invention. Five patent‑law experts were invited to model ten novelty‑related rules using the KR4IPLaw interface. On average, each participant completed the task in seven minutes, demonstrating the usability of SE for domain specialists. The generated LegalRuleML files were successfully loaded into an existing LegalRuleML reasoning engine, which performed consistency checks and simulated rule execution. The OWL‑2 ontologies were inspected in Protégé, where the class hierarchy, object properties, and restrictions matched the intended legal concepts. Quantitatively, the automatic translation achieved a 98 % correctness rate when compared with manually authored LegalRuleML/OWL artifacts, and the error rate was significantly lower than that of a naïve hand‑coding approach.
The discussion acknowledges that KR4IPLaw currently relies on manually curated term and relation dictionaries, which limits scalability across different legal sub‑domains. The authors propose future work that integrates natural‑language‑processing techniques for automatic term extraction and relation learning, thereby reducing the overhead of dictionary maintenance. They also outline plans to extend the framework to other areas of law (contract law, criminal law) and to align the meta‑model with emerging international standards for legal knowledge representation.
In conclusion, the paper demonstrates that Structured English, when coupled with a well‑designed transformation pipeline, can serve as an effective “bridge language” between legal practitioners and knowledge engineers. KR4IPLaw shows that domain experts can author rules in a familiar, computation‑independent syntax while still enabling downstream automated reasoning through formal standards such as LegalRuleML and OWL‑2. This contribution advances both the theory of semi‑formal knowledge representation and its practical applicability in the legal domain, especially for complex fields like intellectual‑property law.