Workshop project 2012-2013
Yannick Maignien MY Consulting Researcher laboratoire TACTIC. Université Paris X Nanterre
Partnership project :
Paris : Société MONDECA, Florence Amardeilh, Directrice R & D TACTIC. Université Paris X Nanterre et MSH Paris Nord (axe Sens Public : Editorialisation)
Berlin : Prof. Dr. Stefan Gradmann Präsident der Deutschen Gesellschaft für Informationswissenschaft und Informationspraxis (DGI) Humboldt-Universität zu Berlin – Berlin School of Library and Information Science
Montreal : Marcello Vitali Rosati, Chaire Cultures numériques, Université de Montreal
Semantic Web, modal logic, digital creation.
Web of data, or Semantic Web, ‘opening the data,” is it intended to facilitate the digital creation?
In recent years, this question has become topical. Currently, the web of data could be described as a proliferation of different ambitions converging towards a common goal: an adaptive reuse of data regardless of the digital origin context and across the Web. One solution is to provide semantized and structured access to Web heterogeneous content through processes of mediation and content aggregation. These processes are based on standardized knowledge representation languages, and provided with reasoning skills. This allows in particular to describe the knowledge used to assess their suitability and thus open up new areas for possible applications.
The amount of data produced on the Web follows an exponential growth for over a decade. The Web is no longer able to be exploited “humanely” in all the quantitative -and especially qualitative- potentialities required by many social and cultural needs. In addition, the tools associated with the Web show their limits in terms of interoperability and information retrieval.
One way of innovation is at the crossroads of three areas: Semantic Web, modal logic and theory of fiction. This path can be defined as the preservation of the “context of validity” of information. Currently, the Semantic Web standard languages do not define precisely the scope of validity of knowledge. To provide seamless access to information aggregation methods or mediation often neglect the context initially associated with the knowledge and without which they no longer have the same semantics. Indeed, some information is valid at some point in the past or the future, others are only if they satisfy certain assumptions, others are just a beliefs. In the disciplines of language and logic, the concept of “context validity” is called “modality” and is the subject of much research in recent decades. Unfortunately, they have rarely crossed the results obtained in both disciplines, and even less in the latest field opened by the Semantic Web.
Regardless of the evolution of the Web, it is an area where issues of reference and context are particularly studied: the theories of fiction, on the narrative discourse, literature and many areas of aesthetic representation.
Indeed, the fiction may be considered as a no-denotation speech, or “metaphorical denotation or non-denotational reference” (Goodman): “According to Frege, fictional statements have meaning (Sinn), but no referent (Bedeutung).”
Factual context, data, may be fictional, if not false in another. The field of creation is fulfilling these changes in perspective.
For the novel the distinctions between author and narrator (Genette) have been particularly studied as well as the ambiguity or distinctions between real and virtual. These distinctions belong to the pragmatic theory as far as they also involve readers, users, players or spectators (Pavel T.), and not just the syntax or semantics of discourse.
But increasingly over the Internet, many statements have meaning, but lose their referential dimension. At least, the search for reference, denotation and the establishment (or not) of verification prove to be increasingly necessary (links to sources, authorship, context, evidence, etc.).
In addition the Internet, as social network, amplifies this pragmatic dimension: it is the application process, consultation, communication and networking operation that must be taken into account in setting or clarify references to the data (conditional terms, etc.)..
The achievements of the theories of fiction and narrative (closely related to the logic of possible worlds) should be worn for the benefit of an advanced Semantic Web
As part of a research on the engineering of knowledge representation, we propose to participate in an exploratory workshop on the theoretical assumptions of these questions, but also on the application and research perspectives opened by relations between Semantic Web and modal logics. The goal of this workshop is to identify ways to overcome the existing borders between the disciplines regarding the issues related to the context of intormation validity, and suggest ways for the development of “vérifiction” tools.
The aim would be to identify in advance the different skills that may come into fruitful cooperation and foster these issues on 4 or 5 sessions. A seminar (“closed” or by invitation), composed with 15 to 20 participants. Each session would take place around a problematic state of the art or presenting application problem.
This workshop is therefore aimed at designers, computer scientists, philosophers, Netarts creators, developers, logicians, etc..
Florence Amardeilh, Olivier Carloni, Charles Teissèdre, (firstname.lastname@example.org), Mondeca R&D
Yannick Maignien ( email@example.com), MY Consulting Chercheur au Laboratoire TACTIC EA 1738 – Traitement et Appropriation des Connaissances par les TIC – Université Paris X -Nanterre
To go further:
Semantic Web and languages…
The Semantic Web relies on three main languages called “standards”: RDF, SPARQL and OWL. RDF is a language that was originally designed to represent the metadata (or annotations) web resources. It can also be seen as a language for knowledge representation in a graphical form (as defined in graph theory). SPARQL is a language for defining queries to query the knowledge defined in RDF. OWL is a language for defining formal ontologies. A formal ontology is a structured vocabulary of “types” with “properties” that have been mathematically defined. These types are then used in a more concrete level to describe factual knowledge.
The Semantic Web is not limited either to the three languages, (although they are overwhelmingly used). There are others such as RQL (other query language), SKOS is a definition of Terminological language resources, and RuleML / RIF / SWRL languages, that are defining rules. Or languages dedicated to specific uses such as FOAF (social networks), Geonames (geographic), SIOC (Web resources), etc. …
Modal logic …
The area of artificial intelligence product for over half a century results on topics as varied as the expressiveness and interoperability of mathematically based languages, the design of automated reasoning systems, etc. The keystone of this sector is the “classic” logic around which revolves a multitude of other languages, including one called “modal logic”. Modal logic has been created in the Sixties to meet the need logicians had to assign a “validity context” to a logical assertion and to take it into account for the deductions. During the previous half century, a variety of modal logics have been developed and applied for specific purposes: temporal logics (future / past), alethic (possibility / necessity), deontic (prohibition / obligation), epistemic (un-contestability of knowledge), counterfactual (challenged), etc..
A modal logic is already present in the semantic web through the OWL (one fragment corresponds to a so-called multi-modal logic). However, it is difficult to use in OWL to define “validity context” because its function is already modal exclusively for representing relationships. But other modal logics could bring to the semantic Web intelligent means to contextualize the validity of knowledge. Including RDF could be extended by a “modal” layer. The use of the method would have some added value in the areas of Web storefront. For example, the inclusion of temporal and deontic modalities would be of major interest in applications related to the legal field. Moreover, the recent availability of public data on the Web and their mutual interconnection also leads to questions about their context of use.
Theory of Fiction …
Computing and logical developments join or intersect theories of fiction, narration theory (G. Genette, U. Eco, T. Pavel, JM. Schaeffer, ML. Ryan, J. Roubaud, etc..) or theory of possible worlds (S. Kripke, N. Goodman, D. Lewis, etc..) They allow to characterize many heuristics, cultural or social productions (controversy, rumors, illusions, simulations, forward-looking assumptions, augmented reality, storytelling, games, science fiction , etc …) which discriminate between real and virtual, or which define the ““mise en scène” or pragmatic devices (author, narrator, reader, spectator, etc..) . On this matter, see the Fabula Website www.fabula.org/
Obviously, these considerations have implications in areas where the status of truth statements is required (History, Humanities, journalism, law, art creation, etc.).
Support From Professor Thomas Pavel, Professor at the University of Chicago and author of lat of essays, including the « Universe of fiction » TSeuil Editor (Paris) :
« I could finally read your project on Verifiction which I find extremely interesting although its general orientation far exceeds my skills. You use the theory of fiction to better discriminate between the virtual and the real. Indeed, the reflection on fiction focused on these issues. Reading the project I thought that basically the theory of fiction have much to learn from your research. For instance, when one talks about “real” in literature, we think of ordinary physical reality of the world. Sherlock Holmes does not belong to it, it is a creature of fiction. If, however, we target the reality at a slightly more abstract level, for example at the social level, Sherlock Holmes is an invented example of an actual profession, the private detective. It would therefore be worthwhile to develop a theory of levels on which reality is targeted (or avoided). Such a range of realities would also be useful for study of political discourse that also play a wide range of levels of reality. »
Thomas Pavel, March 20, 2012 University of Chicago
Many thanks to Tobias Blake for having re-read and correct that article.