Issue 226: P53 and P56

Posted by Wolfgang Schmidle 10/03/2013 
Dear all, 

I am trying to understand P53 and P56. 

About E18: What would be an example of an E18 Physical Thing that doesn't belong to any subclass of E18, which are E19 Physical Object, E24 Physical Man-Made Thing and E26 Physical Feature? Are there things that are E19 and E26 at the same time? 

P53 "has former or current location" has the domain E18, but the fully developed path starts with E19. Take an E18 that is not an E19, for example the cave "Ideon Andron" in Crete (E26). Is it correct that in this case P53 is not a shortcut of "E19 P25i moved by E9 Move P26 moved to / P27 moved from E53"? If so, is it because going though E9 Move wouldn't make sense? Is P53 the only shortcut that can be extended only for a subset of its domain or range? Is a tsunami wave a (moving) feature of an ocean, or not? 

About P56: Is the following correct? "E19 P56 bears feature E26" can always be extended to "E18 P59 has section E53 Place P53i is former or current location of E18". The converse is only true if the domain of P59 is not only E18 but also E19 and the range of P53i is not only E18 but also E26. Since E26 rules out E19, this means that the P53i in the fully developed path of P56 is in itself not a shortcut here. However, P59 is still a sortcut of "P58 has section definition E46 Section Definition P87i identifies". Taken together, "E19 P56 bears feature E26" can be extended to "E19 P58 has section definition E46 Section Definition P87i identifies E53 Place P53i is former or current location of E26". 
 

Posted by Martin 23/03/2013 

On 10/3/2013 5:27 μμ, Wolfgang Schmidle wrote: 
Dear all, 

I am trying to understand P53 and P56. 

About E18: What would be an example of an E18 Physical Thing that doesn't belong to any subclass of E18, which are E19 Physical Object, E24 Physical Man-Made Thing and E26 Physical Feature? Are there things that are E19 and E26 at the same time? 

Well, there are two answers to that question: 
a) If there exist two or more independent "traits" that justify two or more subclasses of a common superclass, only complete knowledge of the world would allow for assessing that no other trait beyond that exists. Since this cannot be proven, we forbid in the CRM to declare complements. Here, you ask, is E19 complement of E26 with respect to E18, or is E18 - (E19 U E26) empty? Answer: who knows? Very often, we had the feeling to describe everything possible, and later a new species appeared. 

b) There are many instances of E24, for which the decision if they belong to E19 or E26 is neither easy nor particularly useful: Take a house built on sediment ground. Typically, it would be immobile, integrated into the ground below and around it, moving it around would need destruction and reconstruction. 

BUT, for the new airport in Spata near Athens, a complete byzantine church was lifted out of the ground, moved out of the planned runway and "planted" nearby. Is it still the same "object"? Once it has in three dimensions clear borders with the air and ground, we could argue, it's indeed an object, not a feature. 

BUT, what if integral parts of the building are cut into rock? Is the temple in Abu Simbel still the same, or a transformation of the original man-made feature into a in principle mobile construct? 

Has a cave house well-defined parts of its own? There rather seems to be a non-discrete spectrum of cases from clearly man-made features to man-made objects, such that the distinction doesn't seem to be helpful for reasoning with CRM data in these cases. This is not a problem, for these cases we have the superclasses in the CRM. 

P53 "has former or current location" has the domain E18, but the fully developed path starts with E19. Take an E18 that is not an E19, for example the cave "Ideon Andron" in Crete (E26). Is it correct that in this case P53 is not a shortcut of "E19 P25i moved by E9 Move P26 moved to / P27 moved from E53"? If so, is it because going though E9 Move wouldn't make sense? Is P53 the only shortcut that can be extended only for a subset of its domain or range? 

Yes. 
The scope note of P53 says"...In the case of immobile objects, the Place would normally correspond to the Place of creation." We leave it deliberately open if there are things that can "survive" a change from immobile to mobile state without changing identity. See above. Without looking up, I believe there are more cases like that. There is an ISSUE here: The scope note could be clearer about that. It actually should say: Things are either there where the have been created (by man or nature) or where they have been moved to (by man or nature, such as meteorites). 

Is a tsunami wave a (moving) feature of an ocean, or not? 

The Tsumami Wave is a Thing, it is physical, but not an E18 at all. It has no "stability of form" required in the scope note in order to identify it in the sense of museum documentation. It belongs to things the CRM does not detail (yet). The Tsunami can sufficiently be described as event for museum purposes and historical research. Museums do not calculate or document wave forms. 

Nobody has followed the tsunami wave on its trip through the ocean. May be such satellite data exist. It's on a long part of its travel a fairly abstract motion pattern. It does not even transport matter. 

About P56: Is the following correct? "E19 P56 bears feature E26" can always be extended to "E18 P59 has section E53 Place P53i is former or current location of E18". 

Yes. 
The converse is only true if the domain of P59 is not only E18 but also E19 and the range of P53i is not only E18 but also E26. Since E26 rules out E19, this means that the P53i in the fully developed path of P56 is in itself not a shortcut here. However, P59 is still a sortcut of "P58 has section definition E46 Section Definition P87i identifies". Taken together, "E19 P56 bears feature E26" can be extended to "E19 P58 has section definition E46 Section Definition P87i identifies E53 Place P53i is former or current location of E26". 
Yes. 
The whole construct may need a critical revision. Purpose of these properties is to allow for reasoning that features travel around with the objects bearing them. This is always true, if we limit the existence of features to at most the existence of the bearing object. However, it may also be true in a relevant sense, if the bearing object is transformed into another one leaving the bearing region intact. It may be false, if the bearing region is cut out. May be the reasoning can only be a default one. To which degree cutting out a region bearing a feature makes the feature ITSELF an object is even more dubious. Also, here the notion of parthood P46 and position P56 blurs. The part-of notion of a feature is that of being integral to the bearing object. Another notion is the component, an object assembled into another object, and hence removable in a non-destructive way. 

If you have concrete examples for your question, we can discuss these issues in CRM-SIG. From a theoretical discussion only, I don't believe we find a good conclusion on a good practice. 

Thanks for the contribution,