Ontology_ ontologies and ontological reasoning 2 what do

							Ontology, ontologies and ontological
reasoning
                      represent?
2: what do ontologies Joost Breuker
                  Leibniz Center for Law
                       University of
                       Amsterdam
Overview
    Reference to `conceptualizations‟
      Knowledge and semantics
      Kinds of conceptualizations
      Representing/expressing conceptualizations
    LKIF-Core
      Towards a top ontology based on common sense
      A cognitive science perspective
      Main categories explained
Conceptualizations and representations
    “…ontology the specification of a
     conceptualization…”
    Confusion about what ontologies are vs what
     they represent
      They are representations, i.e. expressions in some
       medium.
      They represent conceptualizations of terms used in a
       domain
      Barry Smith‟s (2008) illustrates the confusion…
    Cognitive representations
                Representational artifacts




                             Reality
4
    Ontologies are here




5
    Ontologies are here? But where is
    the conceptualization?




6
or here




7
    or here? But is this the conceptualization of a term
    or of an image??




8
    Ontologies do not represent
    concepts in people‟s heads




9
     Ontologies do not represent
     concepts in people‟s heads…but not
     this way…




10
     Like the scientific theories from which they derive, they
     represent universals in reality
     e.g. leg




11
     Like the scientific theories from which they derive, they
     represent universals in reality
     e.g. leg




12
Cleaning up?
    The `scientific‟ theories are conceptualizations
      The leg is an instance of leg
      There are `scientific‟ (eg in anatomy) and common
       sense legs (eg `nice legs‟), …
    Where‟s HERE the ontology?
    Extra misleading due to using 2`analog‟
     representations:
      Screen
      Drawing (via some `conceptualization‟: perception)
    An ontology is a symbolic representation of
     concept (eg leg)
      This representation should capture the meaning…
Having this clarified, there remain new ambiguities
    What kind of conceptualizations do ontologies
     capture?
     1. How we directly understand and communicate the
        world (common sense), or
     2. Models of worlds derived from controlled observation
        (`scientific‟)
     3. (Models from speculation (eg meta-physics))
    Ontologies represent:
     1. (Parts of) knowledge (bases, specifications) (AI)
     2. Repositories of semantics (Semantic Web,
        Computational linguistics)
Semantics and knowledge
    What we know about terms vs what terms mean
     in a particular context (domain, document,
     phrase,..)

    Semantics is the result of applying knowledge to
     data:
      It gives meaning to data (signs)  information
      In the process of understanding, those `properties‟
       of terms are selected that make up a coherent
       `macro-structure‟ (model).
      It is contextualized knowledge
knowledge and semantics               (cf Levelt, “Speaking”, CUP 1993, fig 3.1)


     spatial                                                     semantics
representational                                                 (meaning/sense)
     system


                                                   semantic
                     propositional             representations
                   representational               (preverbal       FORMULATOR
                        system                   messages)


   kinesthetic
   and other
representational
     systems                                                     ontologies?

                                                knowledge
sense vs meaning

   what is this?                                         this is a car



                   • in traffic: a car is a vehicle, moves, transports,…


                   • for the mechanic: a car is a device, has a motor, etc


                   • for a car salesman: a car is a commodity, has a
                   price, a colour, accessories,…


                   • for an insurance inspector: “is
                   this a car or a wreck?”
context dependency of meaning

                            transport
      level of
                                        vehicle
      abstraction
                                                  car
                                                          taxi



                           context dependency

   the more abstract, the less properties, and the less possible
    variation in meaning/sense

   context dependency: views select properties

    levels of ontologies

                                  transport
         level of
                                              vehicle
         abstraction
                                                        car
                                                                    taxi



                                 context dependency
     top, upper, foundational ontologies
        the `primitives‟ on which we build our knowledge (eg space, time,
         object, process, substance, etc.)
     core ontologies:
        some field of practice, discipline (e.g. medicine, law, etc.)
     domain ontologies:
        the domain of interest, e.g. (Dutch) traffic law,
   The more abstract, the more meaning overlaps
    with conceptual knowledge
   Domain ontologies are usually capturing
    `semantics‟: this limits reuse
   Lingua Universalis view:
     Words < -- > Concepts


   Now:
     Words < -- > Meaning < -- > Concepts < -- > Knowledge
                              |
                          Context
What kind of conceptualization?
    Domain ontologies consist usually of a
     specialized terminology and a particular view
     on some world.
    Core ontologies mediate between some `top‟
     ontology and the main concepts in a domain
      Eg: medicin  scientific notions about biological
       processes
      Eg: law  refined terms to interpret `common
       sense‟ events (sense)
    Current top ontologies do not take common
     sense seriously, neither do they commit to
     `real‟ scientific concepts
    For example: space & time
Space in top ontologies: a `scientific‟ perspective
    SUMO, DOLCE, CyC, BFO, etc
      3 dimensions: 3D ontology for objects (continuants)
      Time added:4D ontology for processes (occurrents)




                                             t3
                                    t2

                       t1
Space is viewed as in outer space (mechanics)
    The 3 dimensions are of equal importance, even
     translated in common-sense terms
      Front-back
      Up-down
      Left-right
Looking at space on earth…
Abstracted…
Space in a common-sense ontology
    Space for positions/places
      A horizontal plane on which objects can stand:
           2 eyes: horizontal front-back  depth
           Left-right confusion (look at the mirror)
      A vertical axis: up-down
           Supported by gravity
    Space of objects (extension)
      3D
    Combination:
      We can `see‟ how:
           Objects can `survive‟ in constructions or assemblages
Combining the `scientific‟ with the common sense view…
LKIF-Core: a common sense ontology for law
    Law consists of terms to interpret cases in the
     real (social) world
      Cases are described in common sense terms
      Legal technical terms are usually more strictly and
       explicitly defined common sense concepts
           Eg liability, responsibility, permission, murder, etc.
      Legislation and contracts often contain explicit
       definitions of terms  `ontology‟
      `Aligning‟ two vocabularies
Dependencies between types of legal knowledge




Legal domain
  ontology
LKIF-Core: a common sense ontology for law
    Law consists of terms to interpret cases in the
     real (social) world
      Cases are described in common sense terms
      Legal technical terms are usually more strictly and
       explicitly defined common sense concepts
           Eg liability, responsibility, permission, murder, etc.
      Legislation and contracts often contain explicit
       definitions of terms  `ontology‟
      `Aligning‟ two vocabularies…
…there are exceptional domains where common sense terms cannot be aligned..
Top-down development/KA support
  LKIF-Upper
  upper ontology     physical                      mental                       social
                     concept                       concept                     concept
          physical      physical               mental
          process        object                object
                                     content
                                               intention             role
          action        document      agent                   norm            organization
                                               “anchors”

 LKIF-Core
 legal core ontology
            legal          legal       legal                 legally            judicial
           action          code       person               valid normjudge    organization
                                                normative
                                                 article

           crime       Dutch penal               article of    perpetrator/     criminal
                                     citizen
                       code (WvSR)                 WvSR         accomplish       court
  legal domain ontology:                                                       Is-a
  (Dutch) criminal law                                                        Part-of
Common sense
    By definition: the knowledge that we all share
     (in a culture)
      Test: what can be left unsaid
      `tacit‟ knowledge


    Descartes (Discourse de la Methode):
      “Nobody complains having a lack of common sense”
Nobody?
An evolutionary perspective on acquiring c.s. concepts
     Concepts enable an organism to perceive
      things and events as instances of objects,
      resp. processes
     Cultural evolution:
       Accumulation:`standing on the shoulders of
        giants‟ (Newton)
       Survival: passing the empirical reality test
        (Popper, 72)
            Taxonomy of `survived‟ knowledge species
     Evolutionary psychology
       It all started in the biology
       `instincts‟ (e.g. Pinker, 2008)
Distance sensors & locomotion
    Space: the canvas
      Vertical: gravity (kinesthesis)
      Horizontal: stereo (f/b; l/r)
    Statics:
      Objects at positions
           Mass, matter (substance), extension
    Dynamics: change
      Events with `speeds‟
             Processes (causes of change)
    Complementary view: background/focus
      Hobbs (95, 05)
Predators and prey
    Agents
      Causation and intention (threat)
    Cooperation
      Differentiation (sex)
      (proto-) roles
Enabling reflection: reification by propositions

     spatial
                                               some higher
representational                               mammals
     system


                                          semantic
                     propositional    representations
                   representational      (preverbal       FORMULATOR
                        system          messages)


   kinesthetic
   and other
representational
     systems




                                                           Joost Breuker
                                                        Cicle Aranguren 2005
The mental world:
    `Propositional‟ representation:
      input mode independent
      `interoperability‟
      Reflection


    Mental life as a metaphor of the physical world
      Mental processes and mental objects
The social world:
    Configurations of roles
      Dissociating agent from role
             Analogue: function of physical object (device)
    Role
      Mental object
      Reciprocal relationships
           transactions
      Prescriptions of behavior
           Prediction by a teleological perspective
           Norms
      Roles are associated with positions
LKIF-Core: main `worlds‟
    physical world
    mental world
    roles (= social world)
    abstract world

    occurrences (terms to refer to occurrences)




                                         Joost Breuker
                                        FOIS-2004, Torino
physical world
    basic `natural‟ concepts: energy & matter
    basic defined concepts: physical object &
     process
      both contain mixtures of energy & matter
      processes are changes
           transfer (changing positions)
           changing value (quality; quantity)
           transformation (changing type of process or object)
      types of processes
           mechanics: movement (moving objects & oneself)
           thermo-dynamics: heat exchange, burning, (friction)
           light (radiation)
           chemistry: solving/mixing/cooking substances Joost Breuker
           …
process and object


   mass                                                  force
                            physical
                matter      concept     energy            heat
 substance                part-of                         electricity
               object                  process

     form                heat exchange
                         movement                       transfer
             size
                         radiation                    transformation
                         change-of-substance     change-of-value




                                                     Joost Breuker
Between death, life and mind
    Biology/life:
      Living and moving physical objects: agents
      Agents have minds
            Minds contain (memories, …), intentions
      Processes initiated by agents: actions
    Awareness: communication actions (cf speech
     acts)
    Self awareness: reflection
        Control over reasoning
        Modeling fellow agents
        Modeling discourse
        …human minds…
                                                        Joost Breuker
                                                       FOIS-2004, Torino
the mental world as a metaphor of the physical world
     mappings:
       energy --> emotion|motivation
       matter/substance --> thought/content (information)
       object ---> mental-object (concept,…)
            container ----> mind, memory
       process ---> mental-process (thinking, memorizing, …)
            process --> action
     mind/body `problem‟:
       person has mind; mind is container of mental entities
       action: will as `force‟ (energy to load the intention)



                                               Joost Breuker
                                              FOIS-2004, Torino
roles
       distinguishing between
           role and role taker: e.g. student - person
       roles imply complementary relations
           speaker-hearer, student - teacher
           these `complementary relations‟ explain duty/rights relations in legal
            theories
       roles are behavioural pre-scriptions
           requirements for role taking (cf man taking `mother role‟)
           norms, prescriptions
           role is subclass of mental object
       role performance may be assessed against role
           Bad cook, good cook, …
           violating legal norm
       social organization: part-of structure of roles, defining social
        positions


                                                                Joost Breuker
                                                               FOIS-2004, Torino
abstract concepts
    limited to purely formal, mathematical
     concepts
      evolutionary starting with count-number (cf Lakoff &
       Nunez, 2000), but also point, line, size (geometry)


    the `concrete‟ vs `abstract‟ distinction is
     covered in LRI-Core mainly by `physical‟ vs
     `mental‟
      mental objects: believe, thought, …
           also non formal views on proposition, set, logic, rectangle,
            ..
      mental (epistemological) roles: hypothesis,
       evidence,…
                                                         Joost Breuker
and to be able to talk about occurrences…
    entities = ((instances of) individual objects)
    events and states of entities = (explained by
     processes)
    situations and histories of entities= information
     management (episodic memory)
    causation as the glue between events
    on the canvas of space and time
      spatial positions/areas
      temporal moments/durations
      „now‟ appears to move by the arrow of time:
       existence of objects as trajectories in space/time
                                               Joost Breuker
                                              FOIS-2004, Torino
LKIF-Core development
    Common-sense top-ontology (`LRI-Core‟)
    Collecting terms (plm 200) from legal experts (6)
        Ratings (relevance, abstraction level, common-sense, etc.)
    Middle out approach: identifying clusters  modules




                                                             Joost Breuker
LKIF-Core: levels and modules
You can get LKIF core at
    http://www.estrellaproject.org/lkif-core/
    See also Deliverable 1.4


    Literature:
    Rinke Hoekstra, Joost Breuker, Marcello Di Bello, and Alexander Boer. LKIF core:
     Principled ontology development for the legal domain. In Joost Breuker, Pompeu
     Casanovas, Michel Klein, and Enrico Francesconi, editors, Law, Ontologies and the
     Semantic Web, volume 188 of Frontiers of Artificial Intelligence and Applications.
     IOS Press, Amsterdam, 2009.