1 / 51

Orientation: Spatial and Temporal Ontologies and some relations to Language

Dagstuhl Seminar 10131 Spatial Representation and Reasoning in Language: Ontologies and Logics of Space. Orientation: Spatial and Temporal Ontologies and some relations to Language. John Bateman University of Bremen. Orientation.

elias
Download Presentation

Orientation: Spatial and Temporal Ontologies and some relations to Language

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Dagstuhl Seminar 10131Spatial Representation and Reasoning in Language: Ontologies and Logics of Space Orientation: Spatial and Temporal Ontologies and some relations to Language John Bateman University of Bremen

  2. Orientation • Space, Time and Language:many groups, many disciplines, many methods: too little interaction • Challenges from language: what kinds of phenomena?many standard examples: but what is the response to them? • Approaches and Open Issues

  3. Many areas....

  4. well-formed.eigenfactor.org

  5. Cognition -> ?

  6. Language and Space • natural language and cognitive semantics as part of the long term program to provide a semantics for natural language: • lexical conceptual structure (LCS)Jackendoff 1983; Landau and Jackendoff 1993, Jackendoff 1999 • linguistic details of spatial expressionsTalmy 1983; Langacker1987; Talmy 2006, Tyler and Evans 2003 All references from: Bateman, J. A. (2010), 'Situating spatial language and the role of ontology: issues and outlook', Linguistics and Language Compass. To appear.

  7. Language and Space • spatial language and psycholinguistics:participants are asked to produce responses in situations exhibiting finely controlled spatial and linguistic variationCarlson-Radvansky and Irwin 1993; Tversky 1993; Coventry et al. 1994; Taylor and Tversky 1996; Bloom et al. 1996; Tversky and Lee 1998; Coventry 1998; Carlson 1999; Tversky 1999; Coventry and Garrod 2004a, b • monologue • dialogueSchober 1993; Tverskyet al. 1999; Andonova et al. 2008, Coventry et al. 2009

  8. Language and Space • spatial language usage • anthropological: Levinson 2003; Mark and Turk 2003 • particular languages • English: Herskovits (1986)... • French: Vandeloise1991; Asher and Sablayrolles 1995; Aurnague et al. 2007, • German: : Bierwisch 1988; Lang et al. 1991; Eschenbach, Tenbrink 2007 • spatial expressions across languages • specific language pairs: Slobin 1996; Bowerman1996; Choi-Jonin and Sarda 2007; Tenbrink 2009 • typological aspects: Talmy 1985; Levinson 1996; Kracht 2002; Berthele 2004; Levinson/Wilkins 2006; Hickmann/Robert, 2006 • language plus other modalities: diagrams, gesture, ...Habel, Wachsmuth, Timo Sowa, Tversky • language developmentBowerman broad vs. particular parts of speech

  9. Language and Space • spatial language semantics • natural language semantics and spatial cognition (QSR): Eschenbach 1999, 2005; Zwarts 2005 • ontological characterisation of space:Vieu 1993; Philippe Muller, Bennett 2001, Bennett and Agarwal 2007 • formal natural language semantics:Bierwisch 1988; Pustejovsky 1991; Fong 1997; Zwartsand Winter 2000; Maienborn 2001; Kracht 2002, 2006;Francez and Steedman 2006

  10. Language and Space • spatial language undertaken within specific application fields • GIS • Robotics • relations to sensor data • field potentials:Stopp et al. 1994; Olivier and Tsujii1994; Stopp 1997; Kray and Blocher 1999; Roy et al. 1999; Zimmer et al. 2001;Kelleher and Costello 2009 • GIS-compatible semantics:Rashid et al. 1998; Egenhofer et al. 1998; Mark et al. 2003; Mark and Turk 2003 • embodied language use:Levit and Roy 2006; Mavridis and Roy 2006;Steels and Loetzsch 2009, AmitabhaMukerjee,

  11. Distinct Ontologies where Space is considered

  12. Efforts at Integration, Synergy, Interaction • SPACENET (EU) • Collaborative Research Center for Spatial Cognition at the Universities of Bremen and Freiburg: SFB/TR8 (German: DFG) • Spatial Intelligence and Learning Center: SILC (US: NSF) • ISO working group • SpatialML ... • This Dagstuhl Workshop

  13. Space & Time • space is not a straightforward extension of approaches to time • many ‘spatial’ phenomena already involve time (and change)

  14. SpaceML (Cristani/Cohn: 2002) OntoSpace/DiaSpace • extension of GeoML • allows: • definition of constant spatial objects • definition of spatial categories • definition of spatial regions • definition of variables over the above • relations between entities allows according to: • RCC-8 • qualitative morphological system MC-4 • quantitative distances • orientation calculi • cardinal relations Intended as representation of interpreted spatial language, rather than language contribution itself

  15. OntoSpace/DiaSpace From language to world… • the degree to which language ‘intervenes’ between us and the world is often under-estimated • there are often assumptions made that some account of language is an account of the world… • i.e., language = ontology (if you do it right)

  16. OntoSpace/DiaSpace Examples of this assumption • straightforward linking of linguistically-motivated organisations and domain/context information • linking linguistic ontologies to domain/world ontologies • Sumo : WordNet • DOLCE : WordNet (OntoWordNet) • (Penman) Upper Model : Domain models • Conceptual hierarchies taken to contain both “linguistic” and “non-linguistic” information: • “we do not think that a clean-cut distinction between these types of information is possible” (VerbMobil Report 29)

  17. But: there are problems… “ What guarantee do we have that by suitably disentangling our linguistic practices we could get closer to the truth? … the view that elements of the way we represent the world linguistically—at the level of logical form if not in ordinary discourse—line up with elements of the world itself. And that view cries for justification; it rests on the dubious presupposition that the lining up is unique when, in fact, the very possibility of genuine ontological disputes suggests the opposite. ” Varzi 2007 “From language to ontology”

  18. But: problems… “ … there is no straight analytic path from language to ontology. … There is, in fact, no way of telling what sorts of things there are given the sorts of things we say. ” Varzi 2007 “From language to ontology”

  19. Varzi’s solution… “ There is, in fact, no way of telling what sorts of things there are given the sorts of things we say. But neither is there a complete gap between our words and the world out there. It’s just that that bridge must be built from below, as it were:ontology comes first, and depending on what we think there is, we must attach a meaning to what we say.Going the other way around is wishful thinking. ” Varzi 2007 “From language to ontology”

  20. But… this is also problematic! • it makes the contrary assumption that finding about how the world is will necessarily tell us how language is treating that world …

  21. Could this work?… “ When we began this work, we believed we could proceed directly to formulate a general logical theory of the concept of place. However, we soon found that the huge variety of different ways in which place enters language made it impossible to achieve a simple theory that covered all these modes. Thus we were driven to a detailed analysis of the many linguistic expressions of place concepts and their semantic content. ” Bennett / Agarwal 2007

  22. Could this work?… “ When we began this work, we believed we could proceed directly to formulate a general logical theory of the concept of place. However, we soon found that the huge variety of different ways in which place enters language made it impossible to achieve a simple theory that covered all these modes. Thus we were driven to a detailed analysis of the many linguistic expressions of place concepts and their semantic content. ” Bennett / Agarwal 2007

  23. Developmental evidence… “ Among students of language acquisition there has been a strong tendency to equate `semantic structure' directly with `conceptual structure'--- … But … the meanings of morphemes and often of larger constructions represent a highly structured and conventionalized layer of organization, different in different languages... ” (Bowerman, 1999)

  24. The complexity and flexibility of spatial language

  25. What does ‘on’ mean? Herskovits (1986)

  26. Herskovits (1986) What does ‘in’ mean?

  27. what does ‘in’ mean? Herskovits (1986:125) “The cat is in the table”

  28. what does ‘in’ mean? The potato is in the bowl Herskovits (1986)

  29. Uses of ‘in’: Herskovits (1986:149) • spatial entity in container • gap/object “embedded” in physical object • physical object “in the air” • physical object in outline of another, or of a group of objects • spatial entity in part of space or environment • person in clothing • spatial entity in area • physical object in a roadway • person in an institution • participant in institution

  30. Bennett / Agarwal (2007) • “I am in Bruxelles” • P(footprint(ext(I)), footprint (ext(Bruxelles))

  31. OntoSpace/DiaSpace Herskovits (1986:88)

  32. Herskovits’ proposal polysemy Sense shifts Ideal Meaning ‘almost true’ tolerance

  33. What does ‘left’ mean? • without context: 90°angle • with road crossing or curve: street’s angle • with object localisation: towards object in region

  34. What does ‘between’ mean?

  35. I5-[DiaSpace] I1-[OntoSpace]

  36. Path between ‘physical’ world and interaction is long veer right veer right take the second exit fork right Klippel, Hansen, Davies, Winter (2005)

  37. OntoSpace/DiaSpace ‘Linguistic’ spatial position dependent on purpose • They are sitting in front of the TV • They are sitting in front of the fridge • They are sitting in front of the post office. • Plant the flowers in front of the house • Park the car in front of the house • They built a new apartment building right in front of the house

  38. I5-[DiaSpace] I1-[OntoSpace] Schematization / Classification A: And this train runs on to Victoria. B: To Waterloo, dear. A: [pause] Well, same thing. B: Well, it’s a little different... A: It’s a London station. And you’ve got to get across... Overheard dialogue; Eurostar. 28.4.2004elderly couple on their way somewhere via Liverpool St.

  39. Schematization • Ontological Diversity • routes • oriented paths • linguistic categorizations • spatial dialogue • reference frames • task dependence

  40. OntoSpace/DiaSpace Inclusion in convex hull Casati & Varzi (1999) Parts and places, p139

  41. OntoSpace/DiaSpace “It is apparent that these cases reveal the limits of the approach insofar as it is purely geometric: a full account calls for a step into other territories where pragmatics, or functional and causal factors at large, must be taken into account.” Casati & Varzi (1999) Parts and places, p140

  42. Functional effects Coventry, Garrod and others

  43. OntoSpace/DiaSpace “From a different perspective one could argue that the applicability of “along” depends on the intention of the producer. By using it instead of the more unmarked case of “past”, a pragmatic goal is achieved such as making sure that the listener gets to see a certain sight, or does not get lost” “However, modelling intentions is quite complicated as it would require the inclusion of, e.g., an explicit user model, a dialog history, etc.” Kray/Baus/Zimmer/Speiser/Krüger (Saarbrücken)

  44. In short… • It is by no means straightforward to go from a ‘geometric’ model of the world • to an appropriate linguistic utterance.

  45. Consequence… • … of starting from accounts of the world and working back to language is that language ends up looking extremely vague

  46. Relating calculi and language • Egenhofer and colleagues

  47. Essentials • Language radically and systematically underspecifies spatial relationships in ways that make a direct linking to spatial representations problematic • If we can capture what language is bringing to the spatial representation, we will be in a far better state to untangle the complexity…

  48. Representations of Space ontology Foundational Ontologies QualitativeSpatialReasoning + Representation Linguistics physical mathematical Geometry R3 BFO DOLCE GFO RCC DC OPRA 9+ GUM-Space ? ? ? ? ? ? ‘alignment’

  49. Many perspectives on ‘reality’: many ontologies space-1 event time space-2 event Ontologically diverse

More Related