Dear All,
Ed raises a couple of good test cases.
[MW>] Ed wrote
I think John Y’s point below is well taken. I would say, however, that “pure abstractions” like ‘classifications’ or ‘sets’ don’t have a lifetime or any ‘spatial’ aspect per se. By definition, they don’t change, ever, and they have no spatial properties. They exist from the moment they are conceived, but it *usually* makes no difference whether you consider such things to have existed a priori, even though they are later added to the universe of reference. That said, when you have a concept in your ontology like ‘class of class’, its instances, which are nominally pure abstractions, become ‘objects’ in the universe of *discourse*, and they have a real lifetime that begins when they are identified as instances/classes of interest.
[MW>] Google did not give me anything useful when I searched for the distinction between Universe of Discourse and Universe of Reference. I understand from you that Universe of Reference are the predicates, and Universe of Discourse are the objects predicates apply to.
I do not accept that becoming an object in the universe of discourse necessarily means they have a real lifetime, even if they are a class, just because they are instances of some class, though I accept that this is a choice you can make.
Can you explain please why choosing that classes in the UoD necessarily have a real lifetime?
For example, there may be standard diagnostic codes for various diseases, and when a new virus is identified, you may have a new ‘class’ of cases, and it might result in a segregation of previous cases that were previously “mis-classified”. >From the point of view of temporal relationships, no change occurred in the temporal part of the case that was classified, and no change occurred in the necessary and sufficient conditions that were taken to characterize the earlier class, but the *name* of the earlier class is now associated with more refined conditions.
[MW>] Surely you could alternatively choose to manage the change in name of particular classes that have extensional identity. For one period a community uses a particular name for one class, and during another period it used the same name for another class. It seems to me you can choose to manage the temporal usage of names, rather than using versioning of temporal classes if you did not wish to admit temporal classes to your ontology.
The managed classification is a versioned object (a thing with temporal parts) such that each version (temporal part) is associated with an invariant ‘class’ defined by a specific set of properties. The ‘misclassified’ case *is* an instance of the class associated with the classification (name) at that time, but it *is not* an instance of the class associated with that classification at a later time. (It is difficult to find anyone in the biomedical communities who thinks that classification is truly rigid.)
[MW>] Or you could say that during one time period you (a community) were interested in one set of classes and at another time you were interested in another set of classes, where all the classes are extensional.
This is a gotcha. Many modelers believe that those things are classes. Being written for some implementation purpose, our ontologies often contain the statement that they are. So, John’s point is actually broader in its impact than one might think. The question is always what you *take to be* invariant. If the ontology takes such a classification to be invariant, then it does not apply to the world in which the classification can be refined or revised. But it may well apply to the world (use) for which it was intended.
[MW>] It seems to me that there is more than one way you can skin this particular cat, and the choice you make will likely depend on the ontological commitments you are prepared to make.
Regards
Matthew West
http://www.matthew-west.org.uk
+44 750 338 5279
In a similar way, if the ‘endurant’ ME in Cory’s view is the way you choose to make the ontology, and it serves your needs, who is to argue? But if I see ME as the state in which I exist in a historical context, it is just another temporal state. There was no ME (state) in 1900.
-Ed
--
Edward J. Barkmeyer Email: edbark@xxxxxxxx
National Institute of Standards & Technology
Systems Integration Division
100 Bureau Drive, Stop 8263 Work: +1 301-975-3528
Gaithersburg, MD 20899-8263 Mobile: +1 240-672-5800
"The opinions expressed above do not reflect consensus of NIST,
and have not been reviewed by any Government authority."
It does seem to me that space and time are general concepts that can apply to almost any other concept in an ontology. Would process and event be sufficient for us to use to associate time or space with all other concepts that have this need in your ontology?
For the medical record example,
Would there be an case where the medical exam instantiated an Event, which resulted in a Description of the Event. Then if Event included Time and Space this could be specialized as another concept Called MedicalExam. A Process could be defined for a Diagnostic and Treatment Protocol based on the results of the Exam.
Is the temporal information associated with the Event sufficient for al of the temporal elements, probably not since we need the age of the patient at the time of the Event. There may be other temporal information associated with the exam itself, such as whether there was fasting for 24 hours prior to the event, for a colonoscopy. How about the amount of time spent by the medical practitioner for billing purposes.
This is where the context of the ontology provides relevant insights about which temporal elements are of relevance, for example a legal court may be interested in the age of the device used in an examination and date/times of scheduled maintenance and compliance if there was a malpractice suit. This type of information relationship would not necessarily be discoverable except by implication of location and date/time values.
I can see where linking could enable relationships for unintended applications of an ontology. From my perspective the ontology has an intended pragmatic purpose and has within itself concepts and relations to enable capture and representation of knowledge for that purpose. Other temporal and spatial relationships outside that purpose may not be discoverable or linkable?
On Wed, Feb 19, 2014 at 2:08 PM, Cory Casanave <cory-c@xxxxxxxxxxxxxxx> wrote:
John,
Of course 3d Ontologies are important, but I don't see how this is or could be a conflict. The model of a thing over its lifetime or a snapshot/phase of a thing are different. There should be a clear relationship between the lifetime and the snapshot that is compatible with both approaches. What seems to be missing in many ontologies (or DBMS) is the explicit statement of which one is being represented.
As for " data structures, computation, and message passing are independent of any assumptions about space and time ", I don't see how this could be the case. Those structures, messages, etc. are about something - either across a limited timeframe or over all time. It not being explicit does not make it "independent".
E.g. The weight taken at my last medical exam is a measurement of a snapshot of me at that time - my medical record can contain many such measurements. The medical record is intended to represent facts about ME, over all time.
-Cory
> Sent: Wednesday, February 19, 2014 11:36 AM
> To: Ontology Summit 2012 discussion
> Subject: Re: [ontology-summit] The tools are not the problem (yet)
>
> Ed,
>
> Given this rare state of agreement (copy below), I'd like to add one more point.
>
> I also agree with Matthew that a 4D ontology for physical states and processes
> is valuable. But we also need to support the many ontologies based on a 3D+1
> coordinate system.
>
> With an abstract specification of states and processes, the spec's for data
> structures, computation, and message passing are independent of any
> assumptions about space and time.
>
> John
>
> -------- Original Message --------
> Subject: Re: [ontology-summit] The tools are not the problem (yet)
> Date: Wed, 19 Feb 2014 15:41:30 +0000
> From: Barkmeyer, Edward J <edward.barkmeyer@xxxxxxxx>
> Reply-To: Ontology Summit 2014 discussion <ontology-
> summit@xxxxxxxxxxxxxxxx>
> To: Ontology Summit 2014 discussion <ontology-summit@xxxxxxxxxxxxxxxx>
>
> John,
>
> I agree with your position. As I recall, the basic model in ISO 18629 the
> "Process Specification Language" (Gruninger et al.) is just such an abstraction,
> in which 'time' per se is at best implicit. Additional modules of that ontology
> add 'time' concepts.
>
> I am also a fan of Ed Zalta's handling of 'states of affairs' and their relationship
> to propositions and truth values, and that model of states/events/activities is
> also independent of 'time' concepts.
>
> -Ed
>
> > -----Original Message-----
> > From: ontology-summit-bounces@xxxxxxxxxxxxxxxx [mailto:ontology-
> > summit-bounces@xxxxxxxxxxxxxxxx] On Behalf Of John F Sowa
> > Sent: Wednesday, February 19, 2014 2:12 AM
> > To: ontology-summit@xxxxxxxxxxxxxxxx
> > Subject: Re: [ontology-summit] The tools are not the problem (yet)
> >
> > On 2/18/2014 7:00 PM, Barkmeyer, Edward J wrote:
> > > As I now understand it, in your model, the universe effectively
> > > consists of states and abstractions...
> > > The ontological commitments are clear.
> >
> > I agree.
> >
> > But the point I was trying to make is that Matthew's ontology defines
> > states in terms of a 4D universe.
> >
> > It is possible (and I believe preferable) to specify states and
> > processes as abstractions. There is no need to assume anything about
> > a physical time or space (4D, 3D, or whatever) in order to define a
> > Turing machine, a finite-state machine, or a Petri net.
> >
> > If you define states and processes abstractly, you can use them to
> > specify computations, data structures, etc., independently of anything
> physical.
> >
> > If you do so, you can formalize communications among systems in a
> > purely abstract way -- independent of any assumptions about any
> > ontology of physics, space, time, matter, etc.
> >
> > John
>
> ________________________________________________________________
> _
> Msg Archives: http://ontolog.cim3.net/forum/ontology-summit/
> Subscribe/Config: http://ontolog.cim3.net/mailman/listinfo/ontology-summit/
> Unsubscribe: mailto:ontology-summit-leave@xxxxxxxxxxxxxxxx
> Community Files: http://ontolog.cim3.net/file/work/OntologySummit2014/
> Community Wiki: http://ontolog.cim3.net/cgi-bin/wiki.pl?OntologySummit2014
> Community Portal: http://ontolog.cim3.net/wiki/
_________________________________________________________________
Msg Archives: http://ontolog.cim3.net/forum/ontology-summit/
Subscribe/Config: http://ontolog.cim3.net/mailman/listinfo/ontology-summit/
Unsubscribe: mailto:ontology-summit-leave@xxxxxxxxxxxxxxxx
Community Files: http://ontolog.cim3.net/file/work/OntologySummit2014/
Community Wiki: http://ontolog.cim3.net/cgi-bin/wiki.pl?OntologySummit2014
Community Portal: http://ontolog.cim3.net/wiki/
--
Peace be with you,
John A. Yanosy Jr.