Towards a Reference Library of Upper Ontologies the DOLCE point of view Nicola Guarino Head, Laboratory for Applied Ontology Institute for Cognitive Sciences and Technology, National Research Council Trento, Italy Thanks to all LOA people!

Upper Ontology Summit, March 14-15, Summary 1. Lightweight ontologies vs. axiomatic ontologies 2. Why a Reference Library of (axiomatic) Upper Ontologies 3. What might the library look like 4. How to build such a library: a common framework A common minimal vocabulary (meta-ontology?) A common strategy to elicit the (hidden) assumptions behind each UO Common guidelines to express Ontology Design Rationale Common metrics to compare ontologies 5.How to evaluate the practical utility of the library and of the single modules. 6. (DOLCEs basic choices)

Upper Ontology Summit, March 14-15, Lightweight ontologies vs. axiomatic ontologies different roles of ontologies Lightweight ontologies Intended meaning of terms mostly known in advance Taxonomic reasoning is the main ontology service Limited expressivity On-line reasoning (stringent computational requirements) Axiomatic ontologies Negotiate meaning across different communities Establish consensus about meaning of a new term within a community Explain meaning of a term to somebody new to community Higher expressivity required to express intended meaning Off-line reasoning (only needed once, before cooperation process starts)

Upper Ontology Summit, March 14-15, When are axiomatic ontologies useful? 1. When subtle distinctions are important 2. When recognizing disagreement is important 3. When rigorous referential semantics is important 4. When general abstractions are important 5. When careful explanation and justification of ontological commitment is important 6. When mutual understanding is important.

Why a Reference Library of Upper Ontologies Understand disagreements Maximize agreements Promote interoperability A starting point for building new ontologies A reference point for easy and rigorous comparison among different ontological approaches A common framework for analyzing, harmonizing and integrating existing ontologies and metadata standards

Upper Ontology Summit, March 14-15, The WFOL architecture (WonderWeb FP5 project) Top Bank Law 4D 3D Single Vision Single Module Formal Links Between Visions & Modules Space of ontological choices Space of application areas Mappings with Lexicons

Upper Ontology Summit, March 14-15, A common minimal vocabulary (meta-ontology?) What is an ontology Common terms Property vs. relation Property vs. quality (harder…) Primitive/defined relation Conceptual relation vs. extensional relation …

Upper Ontology Summit, March 14-15, Common strategy to elicit hidden assumptions Systematically explore hidden intra- and inter-categorial relationships How is subprocess related to part? What are the possible relations within processes? Use general issues of formal (philosophical) ontology to elicit the assumptions made (possibly by means of simple NL questions) Common examples (such as the statue and the clay…) Exploit formal meta-properties (OntoClean-like)

Upper Ontology Summit, March 14-15, Formal Ontological Analysis Theory of Essence and Identity Theory of Parts (Mereology) Theory of Wholes Theory of Dependence Theory of Composition and Constitution Theory of Properties and Qualities The basis for a common ontology vocabulary

Upper Ontology Summit, March 14-15, Common guidelines to express ontology design rationale Identify issues List possible alternatives Carefully justify and position the choices made with respect to possible alternatives Basic options should be clearly documented Clear branching points should allow for easy comparison of ontological options Tradeoffs with respect to: Choice of domain Choice of relevant conceptual relations Choice of primitives Choice of axioms

Upper Ontology Summit, March 14-15, Comparing ontolgoies: precision and coverage Low precision, max coverage Less good Low precision, limited coverage WORSE High precision, max coverage Good Max precision, limited coverage BAD

Upper Ontology Summit, March 14-15, Evaluating Upper Ontologies Methodology-oriented evaluation Impact on principles and best practices for conceptual modeling Application-oriented evaluation Impact on the development of domain ontologies Impact on the adaptability of domain ontologies to different domains Cognitive and linguistic evaluation Impact on multilingual linguistic resources Experiments based on linguistic corpora Psychological experiments on cognitive adequacy

Aldo Gangemis contribution Ontology patterns Role of socially constructed entities (D&S)

Upper Ontology Summit, March 14-15, Information objects: a foundation for content description [a semiotic ontology design pattern]

Upper Ontology Summit, March 14-15, Qualities in DOLCE: the basic pattern Color of rose1 Red421Rose1 InheresHas-quale Rose Color Color-space Red-obj Quality Red-region Has-part Quality attributionQuality space q-location

Upper Ontology Summit, March 14-15, Concepts and descriptions Reify social concepts to be able to predicate on them Social concepts and roles as first-class-citizens CN(x): x is a social concept Reify contexts or concept definitions, called here descriptions Deal with the social, relational, and contextual nature of social concepts DS(x): x is a description DF(x,y): the concept x is defined by the description y US(x,y): the concept x is (re)used in the description y Introduce a temporalized classification relation to link concepts to the entities they classify Account for the dynamic behavior of social concepts CF(x,y,t): at the time t, x is classified by the concept y

Upper Ontology Summit, March 14-15, Underlying assumptions Descriptions: are created by intentional agents at the time of their first encoding in an expression of a public language cease to exist when their last physical support ceases to exist have a unique semantic content (different, but semantically equivalent, expressions can be associated to the same description) have an internal structure intimately related to the logical structure of their semantic contents; partially accounted by means of the predicate US Concepts: are statically linked to descriptions: they cannot change their definitions inherit the temporal extension of their definitions

Extra slides

Ontology Ontologies and intended meaning Language L Conceptualization C (relevant invariants across situations: D, ) Intended models I K (L) State of affairs Situations Ontological commitment K Tarskian interpretation I Ontology models I K (L) Models M D (L)

Upper Ontology Summit, March 14-15, I A (L) M D (L) I B (L) Area of false agreemen t! Why precision is important

Upper Ontology Summit, March 14-15, DOLCE a Descriptive Ontology for Linguistic and Cognitive Engineering Strong cognitive/linguistic bias: descriptive (as opposite to prescriptive) attitude Categories mirror cognition, common sense, and the lexical structure of natural language. Emphasis on cognitive invariants Categories as conceptual containers: no deep metaphysical implications Focus on design rationale to allow easy comparison with different ontological options Rigorous, systematic, interdisciplinary approach Rich axiomatization 37 basic categories 7 basic relations 80 axioms, 100 definitions, 20 theorems Rigorous quality criteria Documentation

Upper Ontology Summit, March 14-15, DOLCEs basic taxonomy Endurant Physical Amount of matter Physical object Feature Non-Physical Mental object Social object … Perdurant Static State Process Dynamic Achievement Accomplishment Quality Physical Spatial location … Temporal Temporal location … Abstract Quality region Time region Space region Color region …

Upper Ontology Summit, March 14-15, DOLCE's Basic Ontological Choices Endurants (aka continuants or objects) and Perdurants (aka occurrences or events) distinct categories connected by the relation of participation. Qualities Individual entities inhering in Endurants or Perdurants can live/change with the objects they inhere in Instance of quality kinds, each associated to a Quality Space representing the "values" (qualia) that qualities (of that kind) can assume. Quality Spaces are neither in time nor in space. Multiplicative approach Different Objects/Events can be spatio-temporally co-localized: the relation of constitution is considered.