[Top] [All Lists]

Re: [ontolog-forum] Incompatibilities in 3D to 4D

To: "'[ontolog-forum] '" <ontolog-forum@xxxxxxxxxxxxxxxx>
From: Matthew West <dr.matthew.west@xxxxxxxxx>
Date: Wed, 11 Mar 2009 17:31:02 -0000
Message-id: <49b7f567.0702d00a.63a3.55ad@xxxxxxxxxxxxx>

Dear Ali,




Thank you for your insightful response. I will respond holistically.


One of the purposes of this email is to show(? or explore the idea) that there is not a fundamental incompatibility between the perdurantist and endurantist points of view - at least at the pragmatic level.


I should clarify some terms right now. I lifted "ontological equivalence" from Gibson's phd thesis. I imagine it to mean exactly what it says, an equivalence on the level of things that exist. IMO this is too strong for the 3D-4D debate, since they each offer a fundamentally different perspective on how things exist.


[MW] I agree with you. The two theories assert that different sorts of things exist – spatio-temporally extended, or just spatially extended.


I used the term logical equivalence only in response to the preceding phrase.


[MW] heuristically I would expect logically equivalent to mean that everything that could be said using one theory could be said using the other. I don’t think this is true, so I think this is still too strong.


Admittedly it is a broad term, so let me delimit it by instead using Definably Interpretable. We can say that ontology A is definably interpretable in ontology B, if the concepts of A may be defined using B's lexicon.


[MW] Much the same as I have suggested, with an implicit “all” before concepts. If it were just “some” concepts then you would be with me.


For a more thorough exposition of this type of "equivalence" people might want to refer to:


Brandon Bennet - Relative Definability in Formal Ontologies - FOIS2004




I think a lot of the apparent incompatibilities arise from choices that are orthogonal to whether one is using a 3D or 4D point of view. For example, a 3Dist might have objects that endure but have no notion of time, instead populating their ontology with events which account for the change that an object undergoes. Though even such a position can be mapped into an ontology which does admit time.


[MW] Actually a 4Dist might do the same, since one view of time is one **** thing after another.


Ultimately, I suppose, my goal is to show that this problem, while certainly real in the philosophical realm, does not pose unsurmountable problems in ontological engineering.


[MW] I agree.


Particularly, if we focus on what the axioms of a given ontology are positing, we may establish logical equivalences which enable the interoperability we seek. At times, we may find that we don't have full logical equivalence, but even partial semantic mappings enable a lot of reuse and interoperability.


[MW] Exactly. Where there is no mapping, the lost knowledge is, more or less by definition, irrelevant in the other ontology. By the way I think interoperability is a good word to use for what we are looking for.


Consequently, I would hope we might shift focus more on how people are using the 3D or 4D views to answer specific questions about their domains, and what things they might expect in the reuse.


Imagine there are two biological ontologies, one in 3D and the other in 4D. If the 4Dist wants to reuse concepts defined in the 3D view, how might they extract the relevant information? What types of contextual information are needed to reconstruct an adequate picture? If i am developing a 4D ontology or 3D one, and if I want to be sure I can interoperate with others, is there metadata that i need to keep track of?  I think these types of questions are pretty useful.


[MW] I’m not sure about meta-data, although that certainly comes into play when you are translating between different formalisms (how was that OWL statement rendered in CL). However, I expect that there is a high level mapping between 3D and 4D that can be defined for those things that both theories can say.


Now on some specific points:

[MW]you need to remember that this is written from a 3D perspective. Strictly, 4D can say things that 3D cannot (no concern to a 3Dist since they would not want to say them). So strictly this is a full mapping in one direction and a partial mapping in the other. The other thing I notice here is that the equivalence seems to be presented in terms of stage theory for 4D rather than perdurance theory (Hawley’s terminology). The stage theory version of 4D has an infinite number of temporal slices linked by some sort of “followed by” relation. Perdurance theory allows for objects that a truly extended in time as well as space. The account above does not seem to account for temporal parts that are not temporal stages, but aggregates of them. [/MW]


Could you posit some things that appear to be untranslatable into the other formulation?  


[MW] I will try to remember some. Perhaps Pat H or Chris P could help me out here.


I haven't had time to go through Hawley's book (you would think in this day and age, access to information wouldn't be the limiter...) in detail yet, except through excerpts on google books. Without actual axioms, it's difficult to show that perdurance theories are definably interpretable in endurance theories, but I am fairly confident it is doable. I haven't come across anything that would suggest that it isn't. We would do well to recall that most 3Dists also admit events, which are 4D entities with temporal parts. 


[MW] Of course, and some of them even use events for the lives of objects, so they can really then do anything 4D can do (but of course then why retain the 3D bit?)


[AH]some cl axioms

[MW] I don’t really follow that I’m afraid. Can you give a clue as to what the variables/constants mean, and which is which.[/MW]


What I was trying to do in the above is to place a restriction on the types of relations a 4Dist would want to enforce.

The first snippet was stating that a particular relation was a relation ranging over 4D objects. The second was to enforce that such a relation requires a time argument.

[MW] Ah. But it doesn’t. That is what the 3D relation needs to have (when did it apply). In 4D the relation would simply between the states of the objects for which it was true, and the time element is derived from the start and end time of the state. You would also need to know which objects the states were temporal parts of of course.


I also neglected to include axioms for what "temporal_part" might mean in such a perspective, which is absolutely essential.


[MW] A temporal part is probably best seen as the intersection of a period (or point) in time and an object (for the whole of its life).



For the record, i'm neither a 3Dist nor a 4Dist, i think each perspective has advantages depending on the types of questions and problems we're trying to answer.  


I guess, if there are no serious errors in this email and the last, are there any practical considerations which might hinder such interoperability (aside from the non-existence of axioms thus far)?  Is the 3D-4D debate really an issue for people developing actual ontology applications?

[MW] Probably not at present, because most people developing an ontology pick one or the other at the outset (or both as in BUFO for different sorts of things). The interest comes when you  want to interrelate existing ontologies and/or use them together.[/MW]


Hehe. I was hoping to see how the 3D-4D debate has practical implications for ontology interoperability. I have difficulty pinning down these problems. Maybe it's still too early to be asking these questions...


[MW] Well as long as we are talking about interoperability, I see no reason why this is not achievable. It would make a good research project for someone.




Matthew West                           

Information  Junction

Tel: +44 560 302 3685

Mobile: +44 750 3385279




This email originates from Information Junction Ltd. Registered in England and Wales No. 6632177.

Registered office: 2 Brookside, Meadow Way, Letchworth Garden City, Hertfordshire, SG6 3JE.







(•`'·.¸(`'·.¸(•)¸.·'´)¸.·'´•) .,.,

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
To Post: mailto:ontolog-forum@xxxxxxxxxxxxxxxx    (01)

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