Doug, (01)
> -----Oorspronkelijk bericht-----
> Van: ontolog-forum-bounces@xxxxxxxxxxxxxxxx [mailto:ontolog-forum-
> bounces@xxxxxxxxxxxxxxxx] Namens doug foxvog
> Verzonden: maandag 10 september 2012 20:41
> Aan: '[ontolog-forum] '
> Onderwerp: Re: [ontolog-forum] Universal Basic Semantic Structures
>
> On Thu, September 6, 2012 16:39, Andries van Renssen wrote:
>
> > Why is a school district not physical? In my view it is a physical
> area on
> > earth with an (unspecified) height and depth.
>
> In that case, a school district would have a mass. A rock and a clump
> of
> dirt would be part of the school district. If i dig to plant a bush in
> my
> yard,
> i am making a hole in the school district.
>
> School districts, in my experience, are defined by boundaries -- and
> thus
> are spatial regions -- they are not defined by mass of certain sorts
> that
> happens to be within those boundaries.
>
[AvR] I agree that a (school) district is defined by its boundary. But the
piece of land that is defined by that boundary is nevertheless a physical
object, and it has a mass, although its value is unknown and not of
interest.
I think that the concept 'school district' is a subtype of role, and not a
subtype of physical object. But the role player is a physical object.
Because for a particular piece of land D that has a role that is classified
as a school district, it holds that D is a physical object, because 'piece
of land' is a subtype of physical object.
Thus both are true:
* D <has a role as a> school district
and
* D <is classified as a> piece of land (i.e. a physical object)
The expression 'a school district is a physical object' is a short-cut for
the extended expression 'a school district D is a physical object with a
role as school district'.
The expression 'D <is a> school dictrict' is a synonyms of 'D <has a role as
a> school district'. I call this a 'classification by role', which can also
be expressed as 'D <is classified by role as a> school district. Those three
expressions are equivalent.
The reason why the expression 'I dig a hole in the school district' sounds
odd is: because that expression is a short-cut for 'I dig a hole in the land
that has a role as school district'. (02)
You defined the concept 'spatial region' and Henk defined the concept
'conceptual reality' as concepts that are by definition non-physical.
My impression is that if you would apply the concept 'role' (of a physical
object), then the role player can still be a physical object, although the
kind of role is a subtype of role and not a subtype of physical object. (03)
> > Physical object (and spatial objects) cannot be located in
> themselves,
>
> This depends upon the definition of "located in". In some systems,
> a "located in" predicate is reflexive.
[AvR] Are those predicates reflexive in order to solve a theoretical
position or is it true that physical objects (including also spatial
objects) cannot be located in themselves? Can you give an example a spatial
object that is located in itself? (04)
>
> > but they all can be in (several) locator as well as in located roles,
> although
> > always in different (individual) relations.
>
> I presume you mean relation statements or "relation instances".
[AvR] Yes, as expressions of different facts.
>
> > I am interested in your subtypes of the <being location in> kind of
> > relation.
>
> Some OpenCyc subtypes of <located in> are:
> aboardTransporter alignedCylinderWithin anchorPointInLineSegment
> artifactFoundInLocation basinOfBodyOfWater carriedIn
> cavityInteriorRegionOf cellularNetworkCoversRegion
> completeOuterLayer connected-SheetTransectsAlong
> connectedAlongInside connectedToInside contentOfFreeSpace
> cospatial countryOfAddress countyInState countyOfAddress
> curveOnSurface damsStream ellipsoidalSection
> embeddedCylinderInSheet endPointOfLineSegment equatorOfSpheroid
> equipmentOfFacility eventOccursAt exactlyLocatedAt-Spatial
> exposureToHarmfulSubstanceOccursAt genericPathway-Complete
> genericPathway-Exact greatCircleOfSphere greatEllipseOfEllipsoid
> greatEllipseOfSpheroid groupMembersFoundInLocation
> groupResidesInRegion groundsOfBuilding hotelInRegion in-Among
> in-ContClosed in-ContCompletely in-ContFullOf in-ContGeneric
> in-ContOpen in-Held in-ImmersedFully in-Permeates in-Rooted
> in-Snugly inRegion inRegion-Underspecified infectingOccursAt
> infoPathway-Complete infoPathway-Exact innervates
> linesInside-Skinlike localeOfElection localityOfObject
> locusOfCellularProcess-Cell locusOfCellularProcess-CellPart
> lowestPointInRegion minimumEnvelopeForGeometry
> motionPathway-Complete motionPathway-Exact objectFoundInLocation
> outerLayer packagedIn parallelOfSpheroid pathInsideRegion
> pluggedInto pointOfSale portalPassedThrough processRunningOn
> protrudesInto residenceOfOrganization salesTerritoryOfAddress
> screwedIn situationLocation smallCircleOfSphere spans-
> Bridgelike
> sphericSection spheroidalSection startPointOfLineSegment
> sticksInto
> sticksInto-2D streetSystemOfArea structureInUrbanArea
> suspendedIntaskInsAreaOfOperations teamRepresentsPolity
> terminalPointOfLineSegment tourIsOfRegion trajectory-Complete
> vertexOfGeometricallyDescribableThing vertexOfLineString
> vertexOfPolygon vertexOfPolyhedron waterOfBodyOfWater
> whollyLocatedAt-Spatial worksFoundInStructure
> In contexts in which a GeopoliticalEntity is considered to be a
> GeographicalRegion as well as an Organization more specializations
> are available:
> cityInState eventOccursInCountry hotelInCity majorCityInState
> metropolitanAreaOf placeInCity stateOfAddress territoryOf
>
[AvR] This seems to be an impressive list of subtypes. I will study them in
more detail, because my question is whether they are kinds of relations that
are required or valuable for a formal language as a definition of a rich
semantic expression capability.
Another question is whether they are all true subtypes of <being located
in>, thus whether the taxonomy is pure. Do you also have subtypes of <being
connected to> and <being a boundary of> and <occurring in>? (05)
At first glance there are a number of the above kinds of relations that seem
to be overloaded with information about their role players.
For example:
* InRegion
* hotelInRegion
* hotelInCity
* placeInCity
* cityInState
* majorCityInState
Why only these and why not many, many more similar ones?
And why is e.g. hotelInRegion not a subtype of InRegion?
But why do you think these add semantic expression power to just <is located
in>?
For example, hotelInCity. This appear to be a subtypes that has a constraint
on its roles that the located role shall be played by a hotel and the
locator role shall be played by a city.
What is the additional semantics of 'Hilton hotel A <hotelInCity> New York'
above 'Hilton hotel A <locatedIn> New York'?
when Hilton hotel A <is classified as a> hotel and New York <is classified
as a> city. What is the reason to duplicate or replace these classifications
by (the definition of) the kind of relation?
It is an implied classification that may even conflict with an explicit
classification. (06)
> > The kind of relation <classification of an individual thing by a kind
> of
> > thing> is semantically different from the kind of relation
> <classification
> > of a kind of thing by a meta kind of thing> as the role players are
> > different.
>
> In that each has different restrictions on argument types, yes.
>
> > In the example, the relation <is classified as a> is a phrase for the
> > first kind of relation.
>
> OK. That was not clear in the text that Kingsley presented.
>
> > Furthermore, the statement is that all individual things 'shall be'
> > classified, whereas that is not required for kinds of things.
>
> This is a rule in your system. Kingsley asked "what is not to like"
> about
> this text snippet. The lack of requirement that kinds of things need
> not
> be classified is something that falls in this category since the lack
> of classification limits the type of reasoning that can be performed.
[AvR] The statement that "all individual things 'shall be' classified" is
indeed a rule. Semantically it is not a necessity in Formal English, because
expressions can be interpreted without it. However, I think it is a valuable
rule, because it adds possibilities for verification of the correctness
(consistency) of expressions. Therefore, the rule is intended as a (strong)
recommendation.
For kinds of things there is another rule that makes their explicit
classification superfluous. That is the rules that their definition shall
include the specification of the supertype concept(s). As the highest
supertype is classified as being a kind of thing, all subtypes inherit that
'classification of class'.
Thus there is no limitation on the type of reasoning. (07)
>
> > Kinds of
> > things shall not necessarily be classified, but 'shall be'
> generalized, by
> > being defined as subtypes of their supertype(s).
>
> Another rule in your system.
[AvR] The question is not whether it is in 'my system'. The question is
whether it is a valuable rule in a kind of Formal English.
In my semantic theory kinds of things are defined by their supertype kinds
of things and the specification of the mutual distinctions between the
subtypes. This rule only states that that should be made explicit. If
obeyed, the rule enables additional verification of the correctness of
expressions and provides knowledge and possible requirements by inheritance.
Therefore, it is a strongly recommended rule. My recommendation is that
parsers should give warnings if the rule is not obeyed, but correct
interpretation is still possible.
Thus: why not? (08)
>
> > Therefore, the term 'individual' is an important semantic
> distinction.
>
> Certainly.
>
> > If we eliminate it the semantic precision would be lost.
>
> As a description of your system, yes.
>
> You are referring to your statement:
> >> > This results in a universal basic semantic data structure for the
> >> > expression of facts about individual things."
>
> So the point is that your system provides a structure for expressing
> facts about individual things, but it does not necessarily provide a
> structure for expression of facts about kinds of things.
>
> By eliminating the restriction to facts about individual things,
> more can be expressed. In responding to Kingsley's question,
> i was stating that i would like the more inclusive definitions
> of data structures better.
[AvR] This is a misunderstanding. I clarified that there is a "universal
basic semantic data structure for the expression of facts about individual
things" NEXT TO a similar "universal basic semantic data structure for the
expression of facts about kinds of things".
Those data structures differ, because there are semantic differences
applicable for the interpretation of the relations.
More inclusive definitions would eliminate those semantic differences, as
was illustrated by the distinction between the semantics of 'classification
of individual thing' and 'classification of kind of thing'.
>
> You also stated:
>
> >> > * Each individual thing is classified by a kind of thing,
> >> > because the meaning of a relation between individual things
> >> > can only be interpreted correctly when
> >> > each related individual thing is classified, as well
> >> > as the roles they play and the relation they have.
>
> I wanted to generalize this to apply to relations among kinds of
> things or between individual things and kinds of things. Your
> semantic distinction means that you are not describing these
> other types of relations.
[AvR] You are too fast and should read a bit more. I discussed the relations
between kinds of things (and the relations between individual things and
kinds of things) separately. That is on purpose, because of the semantic
differences.
More general statements may be fine, but the more specific statements are
required for a precise semantic expression capability. (09)
-- Andries
>
> -- doug foxvog
>
> > In the other case of the use of 'individual' the sentence was taken
> out of
> > contexts, because the original text talks about two basic semantic
> > structures, one for facts about individual things and another for
> facts
> > about kinds of things. So also here the term 'individual' marks an
> > essential semantic distinction.
>
> > I agree that a taxonomy is a hierarchical subtype-supertype network.
>
> > I also agree that each individual thing can (in principle) be
> classified
> > by more than one kind of thing.
>
> > With kind regards,
> > Till after my holidays,
> > Andries
> >
> >> -----Oorspronkelijk bericht-----
> >> Van: ontolog-forum-bounces@xxxxxxxxxxxxxxxx [mailto:ontolog-forum-
> >> bounces@xxxxxxxxxxxxxxxx] Namens doug foxvog
> >> Verzonden: donderdag 6 september 2012 7:29
> >> Aan: [ontolog-forum]
> >> Onderwerp: Re: [ontolog-forum] Universal Basic Semantic Structures
> >>
> >> On Wed, September 5, 2012 12:47, Kingsley Idehen wrote:
> >>
> >> > What's not to like about this excerpt:
> >> >
> >> > "In its simplest form, this is a structure that is also supported
> by
> >> > technologies, such as _RDF_ and _OWL_. However, a semantic model
> >> > includes the following semantic extensions that support an
> improved
> >> > computer interpretation of such sentences and an improved
> >> > computerized
> >> > verification of semantic correctness:
> >>
> >> Fine.
> >>
> >> > * Each kind of relation has a modeled definition. Those semantic
> >> > definitions of the relation type includes the definition of the
> >> > required kinds of roles and the allowed kinds of players of such
> roles.
>
> >> Fine.
>
> >> > For
> >> > example, the relation type <is located in> requires a physical
> object
> >> > in a 'locator' role and another physical object in a 'located'
> role.
>
> >> There are many kinds of "is located in" relations which are useful
> to
> >> tease apart. A more useful, more generic, form would require a
> >> spatial object in both the 'locator' and 'located' role. Non-
> physical
> >> spatial objects (such as school districts or police precincts) could
> >> be in either the 'locator' or 'located' role with such a predicate.
>
> >> I would suggest that the example refer to a "spatial object" instead
> >> of a "physical object".
>
> >> > * Each individual thing is classified by a kind of thing, because
> the
> >> > meaning of a relation between individual things can only be
> >> > interpreted
> >> > correctly when each related individual thing is classified, as
> well
> >> > as the roles they play and the relation they have.
>
> >> I would strike the word "individual", since kinds of things (e.g.,
> >> CanusLupus)
> >> can also be classified by kinds of (meta) things (e.g.,
> >> BiologicalSpecies).
>
> >> I would also clarify this by noting that each thing can be
> classified
> >> by one
> >> or more kinds of things.
> >>
> >> > * The kinds of things are defined by at least a relation with
> their
> >> > supertype kinds of things,
> >>
> >> Fine.
> >>
> >> > thus forming a taxonomy of concepts (a
> >>
> >> The word "taxonomy" suggests a tree structure. This should be
> >> clarified to make clear that a directed acyclic graph is a valid
> >> specialization hierarchy.
> >>
> >> > specialization hierarchy, also called a subtype-supertype
> hierarchy).
> >> > This is necessary for the interpretation of the meaning of the
> >> > classifiers (city, tower, and 'is located in', as well as
> 'locator'
> >> and
> >> > 'located').
> >>
> >> > This results in a universal basic semantic data structure for the
> >> > expression of facts about individual things."
> >>
> >> Again, i'd strike the word "individual".
> >>
> >> -- doug foxvog
> >>
> >> > Source: http://www.gellish.net/topics/semantic-modelling.html .
> >> >
> >> > --
> >> >
> >> > Regards,
> >> >
> >> > Kingsley Idehen
> >> > Founder & CEO
> >> > OpenLink Software
> >> > Company Web: http://www.openlinksw.com
> >> > Personal Weblog: http://www.openlinksw.com/blog/~kidehen
> >> > Twitter/Identi.ca handle: @kidehen
> >> > Google+ Profile:
> https://plus.google.com/112399767740508618350/about
> >> > LinkedIn Profile: http://www.linkedin.com/in/kidehen
> >> >
> >> >
> >> >
> >> >
> >> >
> >> >
> >> > _________________________________________________________________
> >> > Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/
> >> > Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-
> >> forum/
> >> > Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
> >> > Shared Files: http://ontolog.cim3.net/file/
> >> > Community Wiki: http://ontolog.cim3.net/wiki/
> >> > To join: http://ontolog.cim3.net/cgi-
> bin/wiki.pl?WikiHomePage#nid1J
> >> >
> >>
> >>
> >>
> >> _________________________________________________________________
> >> Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/
> >> Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-
> forum/
> >> Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
> >> Shared Files: http://ontolog.cim3.net/file/
> >> Community Wiki: http://ontolog.cim3.net/wiki/
> >> To join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J
> >>
> >
> >
>
>
>
> _________________________________________________________________
> Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/
> Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/
> Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
> Shared Files: http://ontolog.cim3.net/file/
> Community Wiki: http://ontolog.cim3.net/wiki/
> To join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J
> (010)
_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/
Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/
Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/
To join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J (011)
|