Hi Francesco, Welcome back. Your post goes through a lot of points.
For the sake of argument I took up the A example. Doing B would have taken me a long time but I think the same exercise could be done. I have nothing against states, but they should come in where they would be needed. In the example proffered, it seems unclear to me what adding states would do but complexify what is already considered a complex standard. So for exercise sake, I express all the statements below in base CRM. I leave it for review, but I think it is a correct translation of all the propositions that expresses what happened and it requires the creation of no new classes or relations. Below the modelling exercise, I have responded to some points in your proposal. I may seem to come out strongly against states. In fact, I’m not. But I think that neither A nor B makes the case for states. > > A. Ownership of a painting > > A1. A person P1 buys a painting in a gallery and receives an official > document proving the legal purchase > > 1. E8 -> p24 -> E22 ‘Painting 2. E8 -> p22 -> E39 ‘P1’ 3. E8 -> p23 - >E39 ->p2 -> E55 ‘Gallery’ 4. E8 -> p4 -> E52 -> p82a -> Some Particular Time > A2. Since then (or since the price was paid) the painting belongs to person > P1 (in the perspective of the owned object) and/or since then, person P1 has > the ownership of the painting (in the perspective of the owner). > Possession continues to pertain to E39 ‘P1’ based on instance of E2 in A1 unless other info introduced in KB Present whereabouts of E22 Painting with no further information remain last known location. Likely in the Gallery in the KB. (we just don’t know right now it’s not documented or state in A1) The instance of E8 instantiated in A1 states precisely that from the time that it begins (A1.4), it is the case that P1 owns the painting (from everyone’s perspective as long as we are under rule of law… does the gallery or his neighbour think differently? There is a contract, acquisition has been effectuated. In a court or in historical accounts this would stand not based on someone’s perspective but on a document or witness). > A3. The painting is delivered at the house of P1. > 1. E9 Move - p25 -> E22 ‘Painting 2. E9 Move -> p27 -> E53 ‘Gallery Location’ 3. E9 -> p26 -> E53 ‘House of P1’ 4. E10 -> P28 -> E39 ->p2 -> E55 ‘Gallery’ 5. E10 -> P29 -> E39 ‘P1’ 6. E10 -> p30 -> E22 “Painting" 7. E10 ->P4 -> E52 ->P82a -> “some real time” > A4. Since then, the painting is in effective possession of P1 (perspective of > the owner). Or the painting is located in P1's house (perspective of the > painting). > > Now, knowing anything further than E9 and E10 instance instantiated in A3 in our KB, it can be inferred, until further info, precisely that ‘A4' until further info added to KB (maximum limit of this relation being existence of house and existence of person I suppose). E10 of A3 precisely states that the physical possession of the object (according to anyone who can find out by entering the house with a warrant or is invited in for tea) lies with E39 from the time of its instantiation. E9 of A3 precisely states that the physical object with regards to its physical location in the world is in the house of P1 from the moment of its instantiation until some other event nullifies that fact. This is a publicly verifiable fact not local to the painting but verifiable through going to P1’s house for tea or breaking in with nefarious intent... > A5. The painting is stolen by person P2 and sold to person P3. > 1. E9/E7 Activity (p2 -> ‘stealing’) -> p7 -> E39 “P2 Thief” 2. E9/E7 Activity (p2 -> ‘stealing’) -> p11 -> E39 “P1” (in the role of victim) 3. E9 Move -> p27 -> House of P1’ 4. E9 -> p26 -> E53 ‘P2 Hideout of P2 Thief' 5. E10 -> P28 -> E39 ‘P1’ 6. E10 -> P29 -> E39 ’P2 Thief’ 7. E10 -> p30 -> E22 “Painting" 8. E10 ->P4 -> E52 ->P82a -> “some real time” (this event would then terminate and provide an end date to the instance of E10 in A3. 9. E10' -> P28 -> E39 ‘P12 Thief’ 10. E10' -> P29 -> E39 ’P3 Dodgy Art Buyer’ 11. E10' -> p30 -> E22 “Painting" 12. E10' ->P4 -> E52 ->P82a -> “some real time” (this event would then terminate and provide an end date to the instance of E10 in A3. 13. E9' Move -> p27 -> Hideout of P2 Thief’ 14. E9' -> p26 -> E53 ‘House of Dodgy Art Dealer' > A6. The painting is in effective possession of person P3 but the legal > ownership by person P1. > this is described by: Ownership of Painting by P1: The instance of E8 occurring in A3 which is still valid and has no contradicting info. Effective Possession by P3 Dodgy Art Buyer: The ‘ instance of E10 in A5 correlated by E9’ in A5 with further details from the instance of previous instance of E9 that fills in the gap of how if got from P1’s house to P3’ house. > A7. P3 brings the painting to restoration > 1. E9 Move - p25 -> E22 ‘Painting 2. E9 Move -> p27 -> E53 ‘Hideout of Thief’ 3. E9 -> p26 -> E53 ‘Conservators Lab’ 4. E10 -> P28 -> E39 ’P3 Dodgy Art Buyer’ 5. E10 -> P29 -> E39 ’Conservator’ 6. E10 -> p30 -> E22 “Painting” 7. E7 -> p7 -> E39 ‘Conservator' 8. E7 -> p16 -> E22 ‘Painting’ Note here how we have also documented how in empirical fact the painting itself has changed hands and location again, by whom and exactly where. > A8. Police finds the painting and brings it back to person P1. > 1. E7 (p2 has type Police Raid) -> p7 -> E74 ‘Police’ 2. E7 -> p12 -> E22 ‘Painting’ 3. 1. E9 Move - p25 -> E22 ‘Painting 4. E9 Move -> p27 -> E53 ‘Conservators Lab’ 5. E9 -> p26 -> E53 ‘P1 House’ 4. E10 -> P28 -> E39 ’Conservator’ 5. E10 -> P29 -> E39 ’P1’ 6. E10 -> p30 -> E22 “Painting” Note here how we have also documented how in empirical fact the painting itself has changed hands and location again, by whom and exactly where. > A9. Person P1 has again effective possession of the painting. > > This is described by instance of E10 instantiated in A8. It is also the case that the legal acquisition which occurred way back in A3 was never contradicted or invalidated. The journey of the painting has been of a change of custody and location thereafter. All these just the empirical facts. > A10. The painting is again restored because P1 thinks the first restoration > was not a good one. > 1. E9 Move - p25 -> E22 ‘Painting 2. E9 Move -> p27 -> E53 ‘P1’ 3. E9 -> p26 -> E53 ‘Conservator 2 Lab’ 4. E10 -> P28 -> E39 ’P1’ 5. E10 -> P29 -> E39 ’Conservator 2’ 6. E10 -> p30 -> E22 “Painting” 7. E7 -> p7 -> E39 ‘Conservator 2' 8. E7 -> p16 -> E22 ‘Painting’ I.e: mutatis mutant the story of A7 but now done by the legal owner (does s/he have better taste?). > A11. Person P1 sells the painting to another collector P4, with all the > needed documents > 1. E8 -> p24 -> E22 ‘Painting 2. E8 -> p22 -> E39 ’P4 New Legit Buyer’ 3. E8 -> p23 - >E39 ‘P1’ 4. E8 -> p4 -> E52 -> p82a -> Some Particular Time 5. E9 Move - p25 -> E22 ‘Painting 6. E9 Move -> p27 -> E53 ‘P1 House’ 7. E9 -> p26 -> E53 ’P4 New Legit Buyer House’ 8. E10 -> P28 -> E39 -> ‘P1’ 9. E10 -> P29 -> E39 ’P4 New Legit Buyer’ 10. E10 -> p30 -> E22 “Painting" 11. E10 ->P4 -> E52 ->P82a -> “some real time” > A12. Person P1 has no ownership nor effective possession more of the painting > but the fact that P1 was former owner —his ownership— remains in historical > records and memories. > > Documented in all of the above. > > Here some points / issues present in the discussion, which I think one should > carefully consider. > > > a. Our aim is to model states of affairs but insofar as we use propositions > to describe them, these propositions have always slightly different > approaches of reality (see A2, A4, etc.), they express slightly different > points of view. The stress is on the phenomenal aspect but the > epistemological aspect can never be totally eluded. > Why would we want to document the feelings of P1 that he was in possession of the painting? Would someone do that in practice? E8 precisely states that it is the case that P1 does in fact have the legal ownership of the thing. > This is particularly true for states of affairs belonging to social life. And > ownership is typically a phenomenon in social life, defined by social rules > and social agreement (cf. B5/B9-11). Absolute states of affairs as they exist > in experimental physics are not always given in social life, especially in > the domain of law. It is not just a matter of disagreement on some aspects of > the state of affair, it is a different way of thinking states of affairs, or > being interested in different aspects of states of affairs (legal ownership > versus effective possession). Furthermore, in historical research some > aspects of social life are often not knowable because of the lack of > historical sources, e.g. B1, but one knows B2. > I did not model B and I think it would need a new relation or two related to the rights discussion in the other thread whereby one would say under which law framework (E29) such a right was acquired. Of course B is much more juicy and interesting. It talks about something people really care about. It is actually of interest (tragic in this case) and not just a banality that two people are at odds over one thing and that they have each what they believe a valid claim to such a thing. In a humble KB, however, shouldn’t we first document just the facts and then use this as the source from which to run interpretation? I have no legal training, but I think that this is how decisions are written, first we gather what is the case and known and verifiable by all sides (otherwise we are not doing history but politics or propaganda or fiction), then in reference to this and with the use of wisdom (which is outside the scope of CRM) and using judgment we come to some conclusions about everyone’s relative right? > b. What is the aim of our modelling: providing interoperability among > different datasets (b1) or detailed reconstruction of states of affairs in > the domain of social, economic, cultural life (b2) ? What do we need to > provide a coherent, sufficiently explicit and reusable dataset in this second > case (b2) ? > I’m not sure that b1 and b2 are incompatible but different scales. b1 gives you a general framework so that if you have no other way to say something, you can say it more vaguely and still document it somehow. b2 allows you in a harmonized manner that allow you to keep speaking the same language as interdisciplinary colleagues at a general level to document with a finer granularity. > In some cases (b1) the CRM core classes and properties are exclusively to be > used, but in some other cases (b2) we need extensions dedicated to specific > domains, making explicit some important states of affairs. > > > The state of affairs we are discussing about is ownership. The examples show > that ownership in social life and history involves many different things: > events that lead to ownership (A1/B1, B1 being possibly more 'passive'), > different aspects of the actual ownership (legal ownership, effective > possession, location of the object), concepts of particular ownership which > remain even after the end of the actual ownership (A12.), general conceptions > about ownership which are valid only in specific political/social contexts > (B5/B9) and which are the social foundation of particular rights (B2). > > It seems difficult to me to use the CRM as such to produce clear and > generally understandable propositions about such kind of complex issues. > > Boldly, I proclaim that my modelling above precisely documents the propositions given in A and each individual aspect (current custody, legal ownership, physical movements). There is nothing stated in A propositions that is not also stated in the model above. > The 'concept' E30 Right has a very concise scope note: > "This class comprises legal privileges concerning material and immaterial > things or their derivatives. These include reproduction and property rights. > > Examples: > copyright held by ISO on ISO/CD 21127 > ownership of the “Mona Lisa” by the Louvre " > > It is useful for merging different datasets on a higher level of abstraction > but probably not for modelling complex social phenomena. > As I stated in the other thread, I share your opinion here. The problem is not the class itself which is intuitively okay I think. It just has an under defined scope note which does not help the user to understand its particular use. It is not meant for many things that we would call a right in natural language. Normally the scope note should then guide the user to understand the way it is intended to be used (the label is a heuristic). In this case, the scope note fails. It should be, as I believe you agree, revised. > The classical solution proposed in the discussion using time-restricted > properties and starting/ending events seems to be very well-suited and > established for the sake of interoperability on a more abstract level (b1) > but not for modelling such detailed descriptions of specific state of > affairs in a way that makes these propositions about social states of affairs > sufficiently explicit and reusable (b2). > > A1, A3, A5, etc., B1, B3, B5, etc. are events and can be modelled as such. > But what is in between (A2, A4, etc.): conceptual objects that have a > time-span on some properties? or 'place holders' for durations of social > phenomena that have no ontological consistency and therefore exist only in > information systems but not in the ontology ? > I can see the need in general for describing that something is known only to have happened from then to then (with no knowledge of events), but in this particular case I don’t see the argument. I don’t see why it would be especially useful to document A2 and A4 etc. separately. Our historian of A knows all the facts and they can state them in a completely coherent and consistent format that could be used and read by any other historian using the base CRM event model. There is nothing more to say about the feeling states of owners, it would just create unnecessary redundancy in the KB. In the modelling of A that I have done above we learn the whole history of that object, all the intrigues, all the by ways and highways it took using 3 basic event classes. It delivers the whole story. Why would you create extra documentation effort to also say P1 thought that he was really in possession of X. Either he was or he wasn’t and this is described in the facts that are expressed by the model above without states. > Again, in my opinion the answer to this question depends on the aim of the > project. If I get it right Robert is interested in shared ownerships (with > different values for the share) and events that happen during ownership > (A9.). Which seems to me to belong more to the aim b2: modelling of more > complex social-economic-legal phenomena. This would mean that one should use > an extension of the CRM. And I would like to share this extension because we > have identical issues in the domain of historical research. > But at the same time the discussion about 'passive activities' could express > the need, probably present in different minds, to have some modelling of > 'static social phenomena' present in the core ontology itself. The issue > about ownership is clearly stated in the official definition of the CRM, > chapter 'Modelling principles', section about Monotonicity : > > "Properties, such as having a part, an owner or a location, may change many > times for a single item during its existence. Stating instances of such > properties for an item in terms of the CRM only means that these properties > existed during some particular time-span. Therefore, one item may have > multiple instances of the same property reflecting an aggregation of these > instances over the time-span of its existence. If more temporal details are > required, the CRM recommends explicitly describing the events of acquiring or > losing such property instances, such as by E9 Move etc. By virtue of this > principle, the CRM achieves monotonicity with respect to an increase of > knowledge about the states of an item at different times, regardless of their > temporal order." > It seems to me that the discussion is not so much (or not only) about a > 'Right' as a conceptual object, therefore a persistent item, but about > ownership, or about effective possession, which are as such social phenomena, > therefore temporal entities, not just concepts. Their identity comes from a > relationship between the owner and the owned object which is situated in time > and space, and which is socially relevant and can be observed, or even > tested, experimented (I can try to steal the Mona Lisa and see if it belongs > to someone...), not necessarily physically but socially. > I think that these problems are worked out in CRM via E8, E9, and E10, not in the sense of providing subtle historical analysis but allowing the documentation of what actually happened. Now it is very interesting and probably necessary in historical research to add on social elements. I can imagine that all E7s and below might happen in relation to some classes of social norms that would include subclasses of mores and laws, but I don’t see that it negates the base system which just takes about mundane movements but rather enriches it by giving additional nodes of information that will support historical interpretation (which surpasses the capability of the KB at this point and is much more pleasant/feasible for humans to read in prose). > In my opinion, this kind of phenomena could be expressed with some kind of > class similar to E3 Condition State (Scope Note: "This class comprises the > states of objects characterised by a certain condition over a time-span") > which is of course restricted at the moment to "prevailing physical condition > of any material object or feature" and therefore hardly usable in this case. > As Martin said at the recent workshop in Plakias, "ownership" (like some > other —ships) is a good candidate for the notion of 'state'. And if we > understand 'state' as a certain condition of a persistent item over a > time-span, it seems to me that this cannot be a 'passive activity' nor an > event (E5) because this class is precisely defined as "change of states in > cultural, social or physical systems". Wouldn't it be quite confusing, and > difficult to explain, to apply the notion of 'change' to an ownership (be it > understood as legal or effective possession) which does not change during > many years? > > Of course, at the moment this issue has to be treated in extensions, outside > the CRM as such, and we will certainly have such kinds of 'states' in the > extension for historical data. But in my opinion the current discussion shows > that there's probably a more general need of deeper study concerning this > kind of social-legal-economic phenomena which exists, and last in time, > independently from the point of view of the observer. > At FORTH we're busily thinking about states based on the examples from Symogi! More soon hopefully for your consideration. Cheers, George
