We are interested in bridging the world of natural language and the world of the semantic web in particular to support multilingual access to the web of data. In this paper we introduce the ULiS project, that aims at designing a pivot-based NLP technique called Universal Linguistic System, 100% using the semantic web formalisms, and being compliant with the Meaning-Text theory. Through the ULiS, a user could interact with an interlingual knowledge base (IKB) in controlled natural language. Linguistic resources themselves are part of a specific IKB: The Universal Lexical Knowledge base (ULK), so that actors may enhance their controlled natural language, through requests in controlled natural language. We describe a basic interaction scenario at the system level, and provide an overview of the architecture of ULiS. We then introduce the core of the ULiS: the interlingual lexical ontology (ILexicOn), in which each interlingual lexical unit class (ILUc) supports the projection of its semantic decomposition on itself. We validate our model with a standalone ILexicOn, and introduce and explain a concise human-readable notation for it.
In this paper we introduce and illustrate the recently begun ULiS project, which aims at redesigning a pivot-based NLP technique, 100% using the semantic web formalisms, and being compliant with the Meaning-Text theory. ULiS stands for Universal Linguistic System, and is a system through which multiple actors could interact with interlingual semantic web knowledge bases in multiple controlled (i.e., restricted and formal) natural languages. Each controlled natural language (dictionary, grammar rules) would be described in a part of a universal linguistic knowledge base (ULK). Besides this, the ULK consists in one specific interlingual knowledge base. Actors could then enhance their controlled natural language through different actions in controlled natural language (e.g., create, describe, modify, merge, or delete lexical units in the dictionaries and grammar rules; connect situational lexical units to interlingual lexical units; add linguistic attributes with their associated rules, etc.).
The aim of this paper is to overview our proposal for the architecture of ULiS, and to introduce and validate the cornerstone of the universal linguistic knowledge base: the interlingual lexical ontology (ILexicOn). 2
The Meaning-Text Theory (MTT). The MTT is a theoretical linguistic framework
for the construction of models of natural language. As such, its goal is to write
systems of explicit rules that express the correspondence between meanings and texts (or
sounds) in various languages
Thus, twelve modules containing transformation rules are used to transcribe
representations of a level into representations of an adjacent level. The main constituent of
the MTT is the dictionary model where lexical units are described, which is called the
Explanatory Combinatorial Dictionary (ECD), and has been the object of many
works on lexical functions, e.g.,
Lexical ontologies and meaning representation languages. Lexical ontologies are
ontologies of lexicalized concepts, widely used to model lexical semantics. Some
have broad coverage but shallow treatment (i.e., with no or little axiomatization) such
as Princeton WordNet
On the other hand, the Universal Networking Language (UNL) is a meaning
representation language, originally designed for pivot techniques Machine Translation. Its
dictionary is an interlingual lexical ontology based on so-called Universal Words ++,
but the lack of argument frames and lexical functions in the UNL dictionary was
pointed out in (Bogulsavsky, 2002; Bogulsavsky, 2005). This is when the idea of an
ECD-compliant interlingual lexical ontology was first mentioned. After the semantic
web formalisms were introduced at the W3C, an attempt to port the UNL to semantic
web formalisms was the topic of the W3C Common Web Language Incubator Group
Positioning of the ULiS project. The lexical resource we propose to develop is an interlingual lexical ontology coupled with a situational (i.e., a generalization of language-specific) lexical ontology, both using semantic web formalisms, and that together form an ECD-compliant dictionary. Benefits of using semantic web formalisms are high as it enables us to construct an axiomatized graph-representation of a lexical ontology, with validation and inference rules. Using SPIN, we propose to include transformation rules directly in an RDF format, on top of the ECD-compliant lexical ontologies, thus obtaining an expert system on linguistics.
The ULiS model is somehow similar to the FunGramKB
Basic Interaction Scenarios with the ULiS The three basic scenarios of ULiS are illustrated on Figure 1 below.
An actor in a situation c inputs some utterance (e.g., in English: "Who killed Mary?") that is first transformed into an RDF situational representation, which undergoes different language-specific process, and which is finally transformed into a CWL-like interlingual representation.
Machine translation. At this stage, depending on the context, the interlingual representation of the utterance may be translated into another utterance in situation d (e.g., in the French situation: "Qui a tué Mary?") through a situational representation (Output1TEXT on Figure 1.
Management of Interlingual Knowledge Bases. Another possibility is that the interlingual representation of the utterance is transformed in a SPARQL request that is applied on an interlingual knowledge base (IKB), which eventually produces an RDF output (e.g., ex:John01). This RDF output is then first transformed into an interlingual representation, then into a situational representation and finally into an output utterance: Output2TEXT on Figure 1 (e.g., "John killed Mary").
Management of the Universal Linguistic Knowledge base. Finally, the third scenario is the human-computing scenario: the SPARQL request is applied on the Universal Linguistic Knowledge base, which is the Interlingual Knowledge Base where the The RDF-World RDF interlingual representations RDF situational representations
IR RDF SRc RDF
SRd RDF whole ULiS is described. Human actors may thus enhance the controlled natural languages through actions stated in controlled natural language.
SPARQL Request
SPARQLRDF + X RDF
IKBRDF
RDF
Output X RDF IR RDF SRd RDF
Thus the interlingual representation format acts as a pivot not only for natural languages, but any interlingual representation may be translated into a SPARQL request, and any RDF graph may be translated to an interlingual representation.
« Interlingual
Lexical
Meta-Ontology » ileximon:ILexicalUnit
« Situational
Lexical
Meta-Ontology » sleximon:SLexicalUnit InterlingualKnowledgeBase
IKB = KB + anchors + transformation rules (ikb:kill,rdfs:range,ikb:Person) (ikb:kill,rdfs:domain,ikb:Person)
pure interlingual features of the ECD ilexicon:Person
englishlexicon:Person espanollexicon:Persona francaislexicon:Personne Other features of the ECD + Links + Transformation rules (ex:John01,ikb:kill,ex:Mary01)
IRs RDF ex:John01
DSynRs SSynRs DMorphRs SMorphRs DPhonRs
SPhonRs SRs RDF The meta-ontology. The interlingual lexical meta-ontology (ILexiMOn) is the schema that the ILexicOn must satisfy to be compliant with the pure semantic features of the Explanatory Combinatorial Dictionary (ECD). It defines meta-classes, uses RDFS and some of OWL full's axioms, and contains ad hoc SPIN validation and inference rules for the ILexicOn and the interlingual semantic representations (ISemRs). The ontology. The interlingual lexical ontology (ILexicOn) is the interlingual dictionary where interlingual lexical unit classes (ILUcs) are formally defined as instances of the ILexicalUnit meta-class from the ILexiMOn. The ILexicOn contains all the pure semantic features of the Explanatory Combinatorial Dictionary (ECD). Any concept expressible in a natural language or a jargon is defined in the ILexicOn that contains: • The formal definitions of the ILUcs (described in section 5.2) • The definitions of interlingual attribute classes (IAtts) (e.g., plural, future, 1st person, indefinite, etc.); • The definitions of the interlingual semantic relations (ISemRels), that are used in the formal definitions of the ILUcs and to construct interlingual semantic representations (ISemRs); • Interlingual lexical functions: every purely-semantic lexical links such as synonymy, and purely-semantic generic constructions such as the lexical function Centr(X), i.e., (the center of X), or Fin(X), i.e., (stop being X . ) The interlingual semantic representations. ISemRs are RDF graphs with nodes being interlingual lexical unit instances (ILUis), and arcs being ISemRels. ILUis may also be instances of IAtts. Arcs are interlingual semantic relations (ISemRels). 4.3
To and from Natural Language facts Situations. Interlingual-based lexical resources consider connecting language specific dictionaries to some interlingual dictionary. We generalize this by using situations (i.e., the situations of understanding and use of some linguistic element).
The situation of a linguistic element is part of the pragmatics of its use: it represents not only the language used (e.g., EN, FR), but also sociolectal marks (e.g., biologists, architects, official, slang, reverential), topolectal marks (e.g., U.S., Canada), chronolectal marks (e.g., old, neologic), and even individual marks (e.g., a particular group of people). The intersection of situations is also a situation (EN-U.S.slang), and so is the union of situations (FR-Canada OR FR-France-old). Architecture of the situational layer. This architecture purposefully mirrors the interlingual layer:
A situational lexical meta-ontology (SLexiMOn) describes the SLexicOn, A situational lexical ontology (SLexicOn), contains all non-purely semantic features of the ECD. A non-exhaustive list is the following: • Definitions of situational lexical unit classes, called SLUcs, by means of a link to an ILUc, which is annotated by a specific situation. • Situational lexical functions such as Instr(X), i.e., the preposition that governs the keyword X and means: (by means of). • Situational attribute classes (e.g., invariable English nouns, French 1st verb group,
German dative, etc.), their associated situations and rules. • Situational relations: relations that link two instances of the SLUcs, thus defining the dependency syntax of the utterance, or the order of the words in an utterance.
Situational representations (SRs). The data consist of situational representations (SRs): RDF graphs having situational lexical unit instances (SLUis) as nodes and situational relations as arcs. A SR thus represents the different representations of the Meaning-Text theory.
Transformation rules. Contrary to the Common Web Language (CWL), where no grammar rules representation is proposed, we plan to introduce transformation rules in the SLexiMOn. Transformation rules form a subclass of the SPIN rules and are attached to a SLUc to define a correspondence between a generic pattern from a representation level, to another pattern at a deeper or to a higher representation level. Thus, each situation may define its own analysis and production grammar, both made of six sets of transformation rules.
Transformation rules may be sorted according to their level of genericity:
transformation rules that are attached to ISemRels, or to IAtts, are less specific than rules
that may be triggered only when a complex ISemR patterns is met; also, rules that
may be triggered in generic situations are less specific than those that may only be
triggered in more specific situations. The important point is that a rule must be
triggered if and only if there is not a more specific rule that can be triggered instead. This
implies that an algorithm different from the simple forward-chaining algorithm must
be proposed. It will be very important to optimize the application of such an algorithm
with a whole set of rules. We therefore plan to construct a Rete network
Finally, a set of generic transformation rules must be designed to ensure that for each situation, every SR is transformable to an ISemR, and that every ISemR is transformable to a SR. When a new situation is introduced (e.g., a new language), this criterion is a priori not met. This is the reason why we suggest the introduction of the universal situation, and transformation rules that produce Notation3-like output. We claim that a small set of rules will suffice to produce and analyze simple controlled natural languages. 4.4
To and from Interlingual Knowledge Bases facts.
Interlingual knowledge bases. The main criterion that an interlingual knowledge base must meet is that any RDF graph inside it must be transformable into an interlingual semantic representation (ISemR). We thus propose to form interlingual knowledge bases by augmenting classic knowledge bases with anchors and transformation rules: • An anchor is a triple that links an RDF resource to an ILUc. For instance, the RDF resource rdfs:Class will be anchored to a specific ILUc ilexicon:RdfClass that formally defines the concept of an RDF class, and that is itself linked to an English SLUc that is a pluralizable noun realized by the string "class"; • The transformation rules are stored in the interlingual knowledge base and form two separated sets of rules: one for producing RDF from an ISemR, the other for producing an ISemR from RDF. Here again, transformation rules may be sorted according to their level of genericity, and the most generic rules must be inhibited when more specific ones can be triggered.
Augmenting classic semantic web formalisms. The output of an ISemR must be a valid SPARQL request, and the output of any RDF graph must be a valid ISemR. This criterion will be satisfied by the introduction of different anchors and generic transformation rules in the classic semantic web vocabularies: RDF, then RDFS, OWL and SPIN, and finally SKOS. Thus an RDF class that has no anchor, e.g., foaf:Person, has a correspondence with an ISemR that itself has a correspondence to the textual representation for the EN situation: "The RDF class foaf:Person". 5 5.1
Modeling Choices in the Interlingual Layer Overview owl:Class owl:ObjectProperty xsd:boolean owl:intersectionOf is-a owl:unionOf owl:propertyChainAxiom owl:hasSelf subClassOf
is-a :ILexicalUnit is-a subClassOf :ILexicalPrimitive :Entity :Person :Mary01 hasEntity is-a :State :Alive :Alive01
subClassOf is-a :ISemanticRelation intersectionOf :onISemanticRelation :allValuesFrom :isObligatory range :hasEntity
true Class/instance
A B A is a subClass of B
B A
C A is the intersection of B and C property A
B A is an instance of B
p A B A is linked to B through property p
RDF/XML document available at URL: http://ns.inria.fr/ulk/2011/06/10/ileximon-core RDF/XML document available at URL: http://ns.inria.fr/ulk/2011/06/10/ilexicon-ex RDF/XML document available at URL: http://ns.inria.fr/ulk/2011/06/10/sems-ex From top to bottom: 1) the semantic web formalisms, with a few OWL classes and properties that are useful for our work; 2) the detailed core-ILexiMOn; 3) an overview of the light standalone ILexicOn; and 4) an overview of data from the interlingual data component. Notice that: i) ILUis from the data are instances of ILUcs described in the ILexicOn, that are themselves instances of the ILexicalUnit meta-classes described in the ILexiMOn; and ii) properties used to link two resources in a layer are described in an upper layer.
ILexicOn – standalone&light
State –(hasEntity)→1.Entity Relation<State –(hasEntity)→1.Entity
–(hasObject)→1.Entity Event –(hasTime)→1.Time Cause<Event –(hasTime)→1.Time –(hasAgent)→1.Person –(hasEvent)→1.Event
Entity Person<Entity
Time<Entity Alive<State –(hasEntity)→1.Person Parent<Relation –(hasEntity)→1.Person
–(hasObject)→1.Person End<Event –(hasTime)→1.Time
–(hasState)→1.State Die<End –(hasTime)→1.Time –(hasState)→1.Alive –(hasEntity)1.Person –(hasState/hasEntity<hasDead)→1.Person Kill<Cause –(hasTime)→1.Time –(hasAgent)→1.Person –(hasEvent)→1.Die –(hasTime)→1.Time
–(hasDead)→1.Person –(hasEvent/hasDead<hasKilled)→1.Person –(hasEvent/hasTime<hasKillTime<hasTime)→1.Time –(hasBeneficiary)→?.Person
Suicide<Kill –(hasKillTime)→1.Time –(hasBeneficiary)→?.Person –(hasAgent)→1.Person –(hasKilled)→1.Person –(hasExperiencer<hasAgent, hasKilled)→1.Person Infanticide<Kill –(hasKillTime)→1.Time –(hasBeneficiary)→?.Person –(hasAgent)→1.Person –(hasKilled)→1.Person – (hasParent)→1.Parent –(hasEntity)→1.Person
–(hasObject)→1.Person –(hasParent/hasObject<hasKillerParent<hasAgent)→1.Person –(hasParent/hasEntity<hasKilledChild<hasKilled)→1.Person ISemR
John kills Mary: Kill: k01 –(hasAgent)→Person: John01
–(hasKilled)→Person: Mary01 definition of the ILUc that is written in its top-left corner. 5.2
The lexicographic definition of lexical units
In the ILexicOn is propose a novel approach to the lexicographic definition of an
ILUc that consists in projecting the minimal semantic decomposition of the ILUc on
the ILUc using Conceptual Participant slots (ConP-slot): the implicit semantic link
that exists between an ILUc L and one of the participants of the minimal semantic
decomposition of L
Interlingual lexical units (classes and instances): ILUcs are instances of the ILexicalUnit meta-class from the ILexiMOn (c.f., Figure 3). They are defined in the ILexicOn (c.f., Figure 4, e.g., Entity, Person, State, Alive, Event, Cause). In our notation, symbol < represents the rdfs:subClassOf axiom that may be used to state inheritance between ILUcs (e.g., Person<Entity, Alive<State, Cause<Event). For instance, The ILUc Person is a sub-class of the ILUc class Entity, and the ILUc Entity is the parent of the ILUc Person. Complex ILUcs may be constructed through owl:intersectionOf and owl:unionOf. Finally, interlingual lexical unit instances (ILUis) are instances of ILUcs and are used in the data component as nodes of the interlingual semantic representations.
Interlingual semantic relations: ISemRels are instances of the ISemRelation metaclass of the ILexiMOn, and thus instances of owl:ObjectProperties. They are introduced in the LexicOn and used in the data to link ILUis (see Figure 3&4). In our notation, symbol < represents the rdfs:subPropertyOf axiom that may be used to define a new ISemRel as being a sub-ISemRel of one or more ISemRels (e.g., hasExperiencer<hasAgent, hasKilled). Symbol / represents the owl:propertyChainAxiom axiom that may also be used to state that a ISemRel is a super-ISemRel of the composition of two or more ISemRels (e.g., hasState/hasEntity<hasDead). These two axioms may be combined to define complex ISemRels (e.g., hasEvent/hasTime<hasKillTime<hasTime).
Interlingual lexical primitives: An ILUc L is a ILPc if and only if it derives from no other ILUc but has at least one ConP-slot. Non- lexical primitives then derive from one or more lexical primitives following the ConP-slot inheritance and introduction principle:
An ILUc L inherits from its parents' ConP-slots, and may also introduce new ConP-slots; One may thus consider only participants that are necessary and sufficient to the minimal projection of L. ILPcs are defined as instances of the ILexicalPrimitive metaclass from the ILexiMOn (c.f., Figure 3). An ILPc must be linked through: i) the onISemanticRelation property to exactly one ISemanticRelation; ii) the allValuesFrom property to exactly one ILexicalUnit; and iii) the isObligatory property to exactly one xsd:boolean.
Conceptual participant slots: In Figure 4, each line with an arrow in the definition of an ILUc represents a conceptual participant slot (ConP-slot) that restricts the use of a specific ISemRel for this ILUc and its descendants. Actually, such a line means that the defined ILUc is a sub-class of an ILPc. For instance, the line State– (hasEntity)→1.Entity states that any instance of the State class is linked exactly once through the hasEntity relation to an instance of the Entity class. Let us focus on the notation used on Figure 4: • Inheritance. ConP-slots may be newly defined (black font, e.g., State–(hasEntity)→1.Entity), fully inherited (grey font, e.g., Relation<State–(hasEntity)→1.Entity) or partially inherited (grey font for the inherited part, e.g., Alive<State–(hasEntity)→1.Person). The ILUc on the right hand side of the line is called the current range of the ConP-slot. • Obligatory vs. optional. A ConP-slot may be obligatory (symbol 1, e.g., Alive<State–(hasEntity)→1.Person) or optional (symbol ?, e.g., Kill<Cause– (hasBeneficiary)→?.Person). When an optional ConP-slot is inherited, it may be restricted to being obligatory. • Domain/range of the ISemRel. As an ISemRel is an rdf:Property, it may restrict its domain and its range i.e., what ILUc the subject (resp. the object) of a triple that involves this ISemRel does belong to. When an ISemRel is underlined, it means that its domain is set to the defined ILUc, and that its range is set to the current ILUc range of the ConP-slot. (e.g., State–(hasEntity)→1.Entity). • ISemRel subproperty and composition axioms. As we stated in section 4.2.2, complex ISemRel may be defined thanks to inheritance and composition. There are benefits in using such ISemRel to qualify a new ConP-slot. In fact, this combined with the maximum cardinality of ConP-slots restricted to 1, imposes the equality of ILUi in the data. We illustrate these inferable equalities by dotted lines on the right of ConP-slots.
The ISemRel inheritance and composition is what enables the projection not only of trees, but also graphs, onto one node. Thus, each ILUc described in the ILexicOn contains the projection of its semantic decomposition graph. We illustrated this on Figure 4 with complex ILUc such as ilexicon:Suicide (the killer is the person killed) and ilexicon:Infanticide (the killer is the parent of the person killed). 6
We introduced a universal linguistic system (ULiS) through which multiple actors could interact with an interlingual knowledge base (IKB) in controlled natural language. We explained an interaction scenario with ULiS, which can serve for machine translation and for multilingual management of interlingual knowledge bases. We then gave an overview of the layers ULiS is made of: the interlingual layer; the situational layer; and an interlingual knowledge base.
The main novelty of our proposal is that the characteristics of each controlled natural language are stored in a specific interlingual knowledge base. Thus, actors could enhance their controlled natural language through requests expressed in controlled natural language.
We introduced and illustrated a novel approach to formally define ILUcs: we make ILUcs support a projection of their semantic decomposition. We introduced a humanreadable notation to represent ILexicOn, and we used this notation to validate our approach with a simple standalone ILexicOn. We thus showed that simple and comc plex ILU s may be formally defined with our novel approach.
We are currently working on the formalization of lexical functions in the ILexicOn and of the SLexicOn, and we are to partly populate our lexical resources with lexical units from other lexical resources such as the French Lexical Network. We finally plan to validate our results by the design and the experimentation of a web-based prototype with a simple interlingual knowledge base (e.g., the "interlingualaugmented" wine ontology), and a few situations based on English and French.