[Top] [All Lists]

Re: [ontolog-forum] What is the role of an upper level ontology?

To: "[ontolog-forum]" <ontolog-forum@xxxxxxxxxxxxxxxx>
From: "doug foxvog" <doug@xxxxxxxxxx>
Date: Thu, 30 May 2013 22:52:00 -0400
Message-id: <40acc96d81d7fcf9c705b9f69f65d7bb.squirrel@xxxxxxxxxxxxxxxxx>
On Tue, May 28, 2013 10:48, Matthew West wrote:
> Dear Doug,
>> On Thu, May 23, 2013 09:25, Matthew West wrote:
>> > doug foxvog wrote:
>> > > Matthew West wrote:
>> >> > MW: I go for strong 4D
>> >> > which has the 4D extent as the identity of any individual (not a
>> >> > set or relationship). So I do not even know what it means to exist
>> >> > if you have a temporal but not a spatial extent.
>> >> At work i have a sick leave account and a vacation leave account.
>> >> Such accounts (as well as bank accounts, credit card accounts) are
>> >> useful to model in an ontology.
>> >> They certainly have temporal existence.  But it seems to me quite a
>> >> stretch to claim that they have a spatial extent.    (01)

>> > MW: The question is what is the root of a day's leave? It is you for
>> > the day in a vacation state.    (02)

>> The root is permission.  The existence of leave in the account is not
>> wedded to any particular day or location (other than near-Earth).
> MW: I think you are making the same mistake as Hans,
> in thinking that what something is, is what creates it.    (03)

I think you guess wrong.    (04)

We were discussing leave accounts, which record permission
by an employer for their employees to take up to a certain
amount of leave.  I.e., to engage in certain types of activity
with certain temporal restrictions.    (05)

Then you asked for "the root of a day's leave".  I provided an answer
regarding "the root" not an answer about what "a day's leave" is.
FWIW, note that the *word* "leave" in this case means "permission"
as in "he gives me leave to X", not "depart".    (06)

> So I repeat, your vacation is you whilst you are on vacation.    (07)

???    (08)

My vacation (not my leave nor my leave account) is an event of
which i am a participant; it is not a 4D slice of me.  As a 4Dist,
you should have no problem modeling events.    (09)

> That is (part of) the fulfilment of some agreement that there is
> some amount of vacation you may take during some period of time.    (010)

The event is part of the fulfillment of the agreement, yes.    (011)

>> > That is a spatio-temporal extent, and any account of it is some
>> > representation of that, or an aggregate of similar objects.    (012)

>> So an account is a representation.  I suppose you would claim that the
>> representation has to be physical.    (013)

> MW: There isn't any other kind.    (014)

>> >> Permission to do something has a temporal extent.  Claiming that
>> >> its spatial extent it the area in which the thing may be done seems
>> >> specious to me.    (015)

>> > MW: Not at all. You only need to bring possible worlds into the
>> > equation
>> So to avoid abstract objects in the ontology, one needs to
>> include an infinite number of possible worlds in the ontology.    (016)

> MW: I have possible worlds to deal with modality. You need some way to
> deal with modality in any practical ontology.    (017)

For ontologies (actually knowledge bases) that need modality, my
preferred method is defining contexts (in Cyc "microtheories").  Sure,
these can be considered to be "possible worlds."    (018)

>> If you allow for more than one "possible worlds", isn't your model 5D?
>> You have continually been stating that your ontology is 4D.    (019)

> MW: If you listen carefully you will hear that I say that my ontology is
> 4D, plus a simple set theory, plus possible worlds.    (020)

You added the possible worlds quite late in the discussion.    (021)

I must note that your definition of 4D is different from most, in that
most do allow for temporal aspatial objects as well as temporal extents
of 1D objects (e.g. the North Pole), and 2D objects (e.g. the Equator).    (022)

>> 1) Everything in a 4D ontology has three spatial dimensions and one
>>     temporal dimension.
>> 2) A 4D worm can be defined for everything in a 4D ontology.
> MW: Not for sets.    (023)

If you grant an exception for sets, do you grant exceptions for numbers?
... for character strings?  ... for planar geometric figures?  ... for
solids?    (024)

Since you are already granting exceptions, why can't you grant one for
temporal, non-spatial entities?    (025)

>> 3) 4D ontologies model possible worlds.
> MW: Each possible world is a 4D extent.
>> 4) #2 + #3: A 4D worm can be defined for every possible world.    (026)

>> How do the 4D worms of two different possible worlds that start at the
>> same 4D point differ?    (027)

> MW: By as much as they do. I allow worlds to branch,
> so if I have a plan for how things will go between today
> and tomorrow, that is a possible world that
> starts from here. However, things do not turn out like that.
> That is another possible world.
> Each shares everything up to the start of the plan, but
> then they branch.    (028)

It sounds like they branch in a 5th dimension.    (029)

>> > to cover the future activity (a spatio-temporal extent) that is the
>> > execution of that permission.    (030)

>> >> A patent is temporal, and is different from the sheet of paper (or
>> >> group of all sheets of paper that document the
>> >> patent).   Claiming that it has a spatial extent of the territory of
>> >> the issuing country also seems a stretch.    (031)

>> > MW: A patent is about (generally) a design - that is universal,    (032)

>> So the design is a physical object that exists in the whole universe?
>> Does it also have a mass?  Or is it the patent whose scope is the whole
>> universe?    (033)

> MW: I said it was a universal,    (034)

I responded to your say a design "is universal", not to it "is a universal".
Sorry for my misinterpretation.    (035)

> i.e. a set. It would be the set of physical
> objects that are made to the design.    (036)

If there are no such physical objects made, that means it is the
empty set.  Oh, i suppose this could be saved by bringing in
possible worlds in which such physical objects are made (and if
it is physically impossible to make such a physical object, possible
worlds with different physical laws would include instances).  So
the cardinality of any design would be a set of cardinality Aleph 1?
Or am i counting wrong, and should it be Aleph 2?    (037)

> There will be a physical representation of the design.    (038)

There are many physical representations of the Eiffel Tower.  The
representations are not the tower.  The physical representations
of the design are not the patent, nor the design.    (039)

>> > and placing restrictions on its use.
>> > Again we need possible worlds and activities, but not objects with
>> > temporal but no spatial boundaries.    (040)

>> And the spatial boundaries of the possible worlds are what?    (041)

> MW: Each possible world has its own spatial boundaries
> that are what they are.    (042)

You have still not specified the 4D physical extent of the patent.    (043)

>> >> What would the physical extent of an ontology be?    (044)

>> > MW: An ontology, is an aggregate of signs of classes, individuals,
>> > relationships, and rules. The signs are spatio-temporal extents of
>> > some sort, even if they are bits in computers.    (045)

And, of course, in human minds.    (046)

>> >> What is the physical extent of an hour, a day, a year, a century be?    (047)

>> > MW: The whole universe. That is actually obvious if you really think
>> > in 4D.    (048)

>> The temporal extent of the day 24 May 2013 is different in California,
>> England, and China.  Relativistic physics recognizes that simultaneity
>> in distant objects is context dependent.
>>  Is your 4D system non-relativistic?    (049)

> MW: You need to take a frame of reference, that is all.    (050)

You need to take a frame of reference that specifies simultaneity
criteria for distant galaxies?    (051)

>> >> A time zone, the Earth, the observable universe, an
>> >> expanding/contracting sphere around the time zone/Earth with its
>> >> radius being in light-units of time, the instantaneous greatest
>> >> distance in time from the specific time to the time of the sphere?    (052)

>> > MW: It actually makes a lot more sense when you add the spatial
>> > dimensions in. I suggest you try drawing a space-time diagram for this
>> > problem.    (053)

>> Note that the relative motion of objects affects the observed temporal
>> differences.
> MW: Again, this is about taking a frame of reference.    (054)

>> >> I accept 4D models of physical objects, situations, and events, but
>> >> also accept the existence of temporal non-physical objects.    (055)

>> > MW: I consider that unnecessary, and therefore to be avoided.    (056)

>> But you don't consider possible worlds as something to be avoided,
>> thus you consider them necessary.  Possible worlds are certainly
>> temporal.  Thus you consider them to be temporal
>> objects which are not non-physical.  So, your
>> position is that possible worlds are temporal physical objects?    (057)

> MW: No. They are spatio-temporal objects, just like the universe we
> inhabit.    (058)

This seems like a "yes", to me.    (059)

>> >> > In the strong 4D that I use, an organization consists of the
>> >> > temporal parts of the people involved in it, whilst they are in an
>> >> > organizational role.    (060)

>> And an organization of organizations -- such as the UN?  Does it consist
>> of physical parts of its member organizations?    (061)

> MW: Essentially it consists of the activity done in its name and those
> doing
> it. Some of that is some parts of its member organizations, but it also
> has
> employees in its own right. Largely the member organizations are
> responsible
> for the charter and objectives of the UN as an organization, (i.e. the
> intentional construction bits again) rather than the execution of those
> agreements.    (062)

You've expanded the definition of organization to include the activity
(i.e., events) done in its name.  So, the Korean War (or to be more
specific, the parts of the war performed by UN (including US) troops)
is-4D part of the UN.  The recent wars on Afghanistan and Iraq are
part (or mostly part) of the USA.    (063)

>> >> Then in any KB using that model, Apple, Inc., didn't have a presence
>> >> in the countries in which it rented post office box numbers, but had
>> >> no employees?    (064)

>> > MW: It is actually an interesting question what exactly is the
>> > constitution of an organization. Is it just the people?
>> > Or does it include what the organization owns/rents as well?
>> > The answer is a matter of law, which may vary from place to place.    (065)

>> I don't think the law concerns itself with the physical constitution of
>> organizations.  It concerns itself with officers of organizations,
>> tangible and non-tangible assets of organizations.  (Oops!  Is there
>> a problem modelling what the International Accounting Standards
>> Board standard 38 (IAS
>> 38)[1] defines as: "an identifiable non-monetary asset without physical
>> substance."?) [1] http://www.iasplus.com/en/standards/standard37    (066)

> MW: This is just a difference of interpretation. Examples they give are:
> Examples of possible intangible assets include:
> computer software
> patents
> copyrights
> motion picture films
> customer lists
> mortgage servicing rights
> licenses
> import quotas
> franchises
> customer and supplier relationships
> marketing rights    (067)

Sounds reasonable to me.    (068)

> We've already covered most of these in principle at least. They seem
> to be either information, or transferrable agreements.    (069)

Neither of which have spatial extent.    (070)

>> > MW: If there was a company set up, and it had a PO box, then there
>> > are officers of the company, and even if they are not resident in the
>> > country, then the company exists (there is not a restraint on
>> > companies to operate only within their country). Whether just a PO box
>> > counts as a "presence"
>> > is interesting. I would look for legal precedence.    (071)

>> So you would equate the legal term "presence" with the physical extent
>> of a company?   Does the law consider the "presence" of a company
>> to follow the business travels of its officers and employees?    (072)

> MW: It is whatever the legal precedence says it is. We are talking of
> things that are socially constructed,    (073)

Correct.    (074)

> so it can go either way depending on what the
> social constructions dictate.    (075)

Except when the social constructions state they are non-tangible.  8)#    (076)

> It can even be different in different countries.    (077)

>> > However, the consequence determines what spatio-temporal
>> > extents count as part of a company, not whether or not a
>> > company is a spatio-temporal extent.    (078)

>> Sure.  Both are just matters of definition.    (079)

> MW: Exactly.    (080)

You just agreed that "whether or not a company is a spatio-temporal
extent" is a matter of definition.    (081)

Perhaps you mis-responded.    (082)

But i agree, and my definition is that an organization is not a spatio-
temporal extent; it merely has a temporal extent.    (083)

>> >> Such a model may be consistent and useful for your purposes.
>> >> But there seems to me to be no reason to impose it
>> >> on everyone who wishes to use anyontology.    (084)

>> > MW: I carefully declare that this is a 4D view of the world.    (085)

>> As mentioned above, it seems to be a 5D view of an infinity of worlds.    (086)

>> > I quite agree that this does not require everyone else to take the
>> > same view. I only argue that it is a valid, compact and rigorous view.    (087)

>> I am not disputing that it can not be consistent with reality if terms
>> such as physical and spatial extent aren't redefined.    (088)

Sorry. A typo. I meant to write "if terms ... *are* redefined."    (089)

>> I haven't seen yet how you
>> collapse the infinitude of possible worlds into the one that happens to
>> continue, or how you differentiate them.    (090)

> MW: I don't need to. They are spatio-temporal extents in their own right,
> and you can either go with Lewis and claim that they are as real as this
> one, or claim that they are imaginary. I don't much care, because the
> effect is the same. ...    (091)

That handles "collapse".  What about "differentiate"?  Are you defining
a few choice "possible worlds" and specified contexts, and asserting
statements in those contexts?  Are the worlds/contexts objects in
their own right that you can assert properties about?  If so, in what
kind of context are such statements made?    (092)

>> >> I'd accept it as an upper-level microtheory, but not as part of an
>> >> upper level ontology designed for general use.    (093)

>> > MW: It is perfectly suitable for general use.    (094)

>> In that most users would have a different mindset, it would
>> not be suitable for them.  An ontology need not proselytize.    (095)

> MW: An ontology is supposed to be about what is, not about a mindset.    (096)

In which case, it wouldn't belong in an ontology at all.  But i grant
that it is fine for an ontology that requires its reasoning.   Many
ontologies are not designed for temporal reasoning, so this provides
a mechanism.  Ontologies that provide other mechanisms, either
4D ontologies that accept aspatial temporal objects, or 3D+1
ontologies (which either accept or don't accept aspatial temporal
objects) would provide other mechanisms.    (097)

An agnostic ontology can select any of the above four types of
upper ontology for temporal reasoning.    (098)

It seems to me that any practical question could receive the same
answer using any of these systems.  They might have different answers
for (imho meaningless) questions such as what is the physical extent
of a conceptual work or organization, but that is not the sort of
question that existing computer systems deal with.    (099)

>> > I have not yet found something
>> > (together with the other elements I mentioned, a basic set theory,
>> > mathematical objects, and possible worlds) that it cannot cover.    (0100)

>> Does  your ontology require geometrical objects (triangles, Platonic
>> solids) to have temporal extent as well as spatial extent?  Or do you
>> accept them as 3D objects?    (0101)

> MW: I'm inclined to accept mathematical objects along with sets as
> universals.    (0102)

OK.  A few more exceptions.    (0103)

>> >> > Note that I have no problem constructing one kind of object out
>> >> > of the temporal parts of another kind (or kinds) of object.
>> >> > In any case, given the nature of space-time, I don't know what
>> >> > it means to exist with a temporal extent, but no spatial extent.
>> >> > I therefore reject abstract individuals as an unnecessary
>> >> > commitment.    (0104)

>> >> Some people would reject a 4D model as an unnecessary
>> >> commitment.    (0105)

>> > MW: If you do not commit to 4D you  have to commit
>> > to something else.    (0106)

>> For some purposes, yes.  For many, no.    (0107)

> MW: OK. Well I guess there are some purposes where you can do
> without either 4D objects or 3D objects plus time,
> but I would not refer to that as only "some" purposes.
> There are very few purposes when you can do without
> individuals (things that exist in space-time) at all.    (0108)

Interesting definition of "individual".  You exclude from the definition
what you are arguing against.    (0109)

> Therefore you have to make some commitment
> as to what sort of objects they are.    (0110)

Exactly why do you have to make such a commitment?    (0111)

> The most popular choices are 3D + time
> (wholly present at each point in time, and passing through time)
> or 4D (extended in both time and space). I consider that to
> be a choice, since they are inconsistent with each other.    (0112)

Note that some systems consider events to be extended in time
(and if they are spatial events, extended in both space and time),
but consider physical objects as 3D+1.    (0113)

There is no inconsistency in such a system.    (0114)

> I would be very
> interested in how you think you can avoid the choice.    (0115)

An atemporal knowledge base needs make no commitment.  E.g.,
* These are a set of parts.
* These are the connections between parts in larger components.    (0116)

>> > It is a choice you have to make, not an unnecessary one.    (0117)

>> A personnel system can assign dates to events.  These assignments
>> can be interpreted one way in 4D, another in 3D+1, both of which
>> would be consistent with "the real world".    (0118)

> MW: Events are almost always interpreted as 4D, so that is not great
> trick.
> Otherwise, the point is that you have to interpret the dates as relating
> to 3D or 4D objects before you do any reasoning, and it is that
> interpretation that ontologises them.    (0119)

It depends upon the sort of reasoning that you are doing.    (0120)

A DB row may state that a person has a certain birth date, a certain
shipping address, a certain standing account, a certain telephone
number, and a certain customer ID.    (0121)

A DB row from a different file might say that a person purchased
certain numbers of three specific products on a certain day of a
certain month of a certain year, and has a certain customer ID.    (0122)

A reasoner need not commit to 3D or 4D objects in order to determine
the address to which the products should be shipped, nor the standing
account to debit.    (0123)

The system need not reason that the birth date applies to the whole
person, that the shipping address, telephone number, and customer
ID apply to different 4D slices of the person, nor that the standing
account must apply to a sub-slice of the slice of the person that the
customer ID applies to.    (0124)

Where is the need for such a system to commit to 3D+1 or 4D?    (0125)

>> > MW: That is because what you have stated includes several short cuts.
>> > If you expand it fully, you could say, the person (for the whole of
>> > their life) that is currently a Red Sox Player (a state of the person
>> > whilst they are a Red Sox Player) is currently father (a temporal part
>> > of the person) to Sue.    (0126)

>> Yes, these are short cuts for the 4Der.  But such translations could be
>> performed when translating from human to 4D.  You need not force
>> everyone to speak 4D.    (0127)

> MW: Of course. And you have to do something similar for 3D too. It is just
> a different translation into a different ontological system.    (0128)

The only translation necessary would be to sprinkle in a few "current" or
"currently", or "now"s.  English allows the timing to be considered
current by default  if the tense accepts it.  "Some Red Sox player is Sue's
father" is sufficient for 3D+1.    (0129)

-- doug foxvog    (0130)

> Regards    (0131)

> Matthew West
> Information  Junction
> Tel: +44 1489 880185
> Mobile: +44 750 3385279
> Skype: dr.matthew.west
> matthew.west@xxxxxxxxxxxxxxxxxxxxxxxxx
> http://www.informationjunction.co.uk/
> https://sites.google.com/site/drmatthewwest/
> This email originates from Information Junction Ltd. Registered in England
> and Wales No. 6632177.
> Registered office: 8 Ennismore Close, Letchworth Garden City,
> Hertfordshire,
> SG6 2SU.
>    (0132)

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    (0133)

<Prev in Thread] Current Thread [Next in Thread>