Post by Wolfgang Schmidle (11 Dec 2024)
Dear All,
The scope note of P4 "has time-span" says:
> More than one instance of E2 Temporal Entity may share a common instance of
> E52 Time-Span only if they come into being and end being due to identical
> declarations or events.
Does anyone have an actual example of this in their data?
Best,
Wolfgang
Post by Akihiro Kameda (12 Dec 2024)
Dear Wolfgang,
Thank you for raising this interesting question.
While working on translating the section on temporal relations, I came
across the following description, which raised a similar question for me. I
would like to share my understanding here, and I invite everyone to point
out any misunderstandings if they notice them.
> For documenting exact time spans that are the result of a declaration
rather than observation, for instance, in order to describe a time span
multiple events may fall into, the property P170 defines time allows for
specifying the time span uniquely by a temporal primitive, rather than by
P81 ongoing throughout or P82 at some time within using an identical time
primitive.
________________________________
1. *Fireworks Aimed for Simultaneous Launches and Bursts (Observed
Simultaneity)*
This is an example that *does not fall under the pattern of sharing a
time-span instance*, as it represents observed simultaneity rather than
simultaneity resulting from identical declarations or events.
Consider the case of two fireworks launched where the launches were
coordinated through mutual visual signals by firework artisans. While this
coordination aimed for simultaneous launches and bursts, due to practical
constraints, the events were not strictly simultaneous.
To express perceived simultaneity, such as when observers state that "the
fireworks occurred at the same time," it is more appropriate to model this
as described simultaneity rather than strict simultaneity usable for
inference. This can be achieved by representing the statement as an E89
Propositional Object, referring to the two time-spans via *P67 refers to*
or *P129 is about*, with supplementary information provided using *P3 has
note*.
________________________________
2. *Laws with Identical Effective Dates (Declared Simultaneity)*
Two laws that took effect on "2020/1/1" and ceased on "2022/12/31" due to
the enactment and repeal of a superseding law illustrate declared
simultaneity. For example, Law A and Law B were declared to be in effect
within the same time-span due to identical legislative declarations. This
is an example of sharing an E52 Time-Span that is entirely based on
explicit declarations, not observations.
________________________________
3. *Exhibition Periods for Paintings A and B (Principally Simultaneous)*
Consider an art exhibition where Painting A and Painting B were displayed.
The time-span during which Painting A was exhibited perfectly matches the
time-span during which Painting B was exhibited. Furthermore, both
time-spans are identical to the overall exhibition period, as defined by
the organizers.
This scenario can be modeled by assigning a shared E52 Time-Span to the
exhibition period and referencing it from the respective temporal entities
(e.g., the display events for Paintings A and B) using *P4 has time-span*.
The simultaneity of these events is defined by the organizational
declaration of the exhibition period, rather than being based on
independent observations.
Other examples of principally simultaneous time-spans include remote
meetings. For instance, a remote meeting held between Company A and Company
B could involve events documented separately by each company. Despite being
recorded independently, the events share a common E52 Time-Span as they
occurred during the same interval but from different locations, with
physical and causal simultaneity.
________________________________
Additionally, in our work with CIDOC CRM's *P170 defines time*, we’ve
provided the following examples to highlight cases where exact time-spans
are defined via declarations rather than observations:
Exhibition Period: "2023/1/1 – 2023/3/31" (E61) defines time "The
exhibition period of Museum Exhibit A" (E52).
Effective Period of a Law: "2020/1/1 – 2022/12/31" (E61) defines time "The
effective period of enacted Law B" (E52).
Cultural Period: "2025/1/1 – 2025/1/3" (E61) defines time "The first three
days of the New Year in Japan (2025)" (E52).
These examples align with the principle that a declared time-span (E52) can
serve as the shared temporal context for multiple events or activities.
I hope these examples are helpful.
Best regards,
Akihiro Kameda
Post by Martin DOerr (13 Dec 2024)
Dear Akihiro,
Thank you for these rich considerations!
- *Fireworks Aimed for Simultaneous Launches and Bursts (Observed Simultaneity)*
This is correct. We may however further distinguish the intended simultaneity from the observed approximate simultaneity. The solution you give appears to me absolutely adequate to describe the intention.
For modelling the observation itself, CRMbase does not offer a direct model for observing events. Only with CRMsci, which is being extended these days by a more elaborate observation model, the perception itself could be documented as an "Observable Situation" within which both launches occur.
Staying within CRMbase, we may use /P82 at some time within /to associate the same time-primitive value. However, if later more precise determination, e.g., by video evaluation, would be available, this seemingly accidental simultaneity would become obsolete.
A more explicit way would be, to use two instances of E13 Attribute Assignment, stating that each of the phenomenal time-spans of the actual launches "/P86 falls within/" the same declarative time-span, i.e. one defined by /P170 defines time. /A respective note can mark the Attribute Assignment as a result of an observation. This would be compatible with the current CRMsci/, /using two instances of S4 Observation./
- 2. *Laws with Identical Effective Dates (Declared Simultaneity)*
Excellent! If you have a published example with reference,we would be glad to insert it into the CRM text.
- *Exhibition Periods for Paintings A and B (Principally Simultaneous)*
Excellent!
I'd argue that this is normally a physical simultaneity as well. If the exhibition does not start or end chaotically, the physical opening and closing of the exhibition space will imply the exhibition periods of both paintings, regardless whether the official declaration of the exhibition period deviates more or less from the physical one, isn't it?
If you have a published example with reference,we would be glad to insert it into the CRM text.
This falls into a general pattern of events being part of other evens for the complete duration of the superevent.
<---wrt "principally simultaneous time-spans include remote meetings"-->
Very interesting! Would be good to discuss in more detail! There is a difficult question to which degree electronic connections/communications make things happening at different locations to be part of the same event. This has been our current standpoint at CRM SIG.
We would be glad to include more examples from your side of the world! Note, that examples need to be real cases that have happened and are published, not characterizations of their kinds only.
Comments?
All the best,
Martin
Post by Wolfgang Schmidle (15 Dec 2024)
Dear Akihiro,
Thank you, that was very interesting! What Martin said, and I have some
additional questions about the difference between sharing a common time-span
and the Allen relation AP22 (formerly P114) "is equal in time to" or its
equivalent P175 + P175i + P184 + P184i.
> strict simultaneity usable for inference
If two temporal entities share a common time-span, AP22 also applies. What
would be an inference where it is necessary/useful to model two temporal
entities explicitly as sharing a common time-span instead of just using AP22?
What would be an example of two temporal entities that do not share a common
time-span but are still AP22 "equal in time to"?
Best,
Wolfgang
Post by Martin Doerr (15 Dec 2024)
Dear Wolfgang,
To my opinion the difference is the following:
a) AP22 is no more in CRMbase. So,if you use only ISO21227, you an only use the shared time-span.
b) If you use CRMarchaeo, AP22 is available.
c) Using AP22 needs one property instance only, a shared time-span needs a class instance and two property instances. So, if the time-span is completely unspecified for either event, AP22 is cheaper to use. If one time-span is already specified by P4, it is the same. d) I support we should declare an FOL inference, that both constructs are equivalent, if there are no other use cases differentiating both.
I would not like to delete AP22 from CRMarchaeo, because it is part of the complete system of Allen's relations.
In CRMbase, using a shared time-span is indeed more economic than using the new time-primitives.
How do you see this?
Best,
Martin
Post by Wolfgang Schmidle (1 April 2025)
Dear All,
Here are a few slides about the new issue "Examples for AP22 is equal in time".
https://docs.google.com/presentation/d/1HFVR2rDgB8CRuKs3cPd4qeI-s_VotdZBw84g2VGQhhs/edit
Best,
Wolfgang
Post by Martin Doerr (1 April 2025)
Dear Wolfgang,
Good analysis!
However, I'd like to point out that in our application we do not require zero or negligible analysis, since the theory we have provided about temporal relations are consistent with a "fuzzy zone" delimiting the intervals. There is rather a notion of "irrelevant fuzziness", such as the exact time passengers go from board, when we compare the whole voyage. We *must *always keep in mind what the purpose is, kind of reasoning is adequate for historical research. It is not only a mathematical exercise. We are not interested in what Allen had in mind.
The density of information is never sufficient to do large-scale reasoning on intervals. If a scholar or scientist gets a result from CRM data and questions such details, she has necessarily the obligation to consult the sources, and not use the CRM data as a surrogate. Too narrow interpretation of tolerable fuzziness increases precision, but we need just the opposite, *increase recall*.
I suggest that this my comment above is transferred into the CRM principles, and be presented for the issue.
Kind regards,
Martin
In the 60th joint meeting of the CIDOC CRM and ISO/TC46/SC4/WG9 & 53rd FRBR/LRMoo SIG, WS gave a brief outline of the issue, namely that he fails to see how the example used for AP22 (The destruction of the Villa Justinian Tempus (E6) is equal in time to the death of Maximus Venderus (E69)) forms a good example.
- For starters, the example is very old, and also very possibly fictitious. It is unlikely that it is plausible.
- The examples of synchroneity provided by AK (for declared, phenomenal, observed) were interesting and helpful for the discussion.
- A suggestion was to replace the example with an example demonstrating phenomenal synchroneity via causality (Antonio Pigafetta recording Magellan’s voyage is equal in time to the voyage (of the Victoria).)
- WS does not think that this is a good example because it allows similar examples to be drafted for different events (like the Titanic’s journey). If it is of historical interest when x person died vs when y person died, assuming that their death was simultaneous with the sinking of the Titanic is not good enough. So synchroneity does not necessarily follow from causality
Discussion points:
- In general, this kind of simultaneity seems to work only with declarative events (typically marriage legally begins simultaneously or ends simultaneously in a divorce; two paintings that are part of an exhibition can be claimed to be on display for the same amount of time, as long as the exhibition is in place). So, it works better with speech acts than with an observed phenomenal reality.
It is an interesting topic, could be pursued further. And the examples always need revising.
Bern, April 2025