Issue 459: modelling principles - introduction text for crm - version 7.0
In the 45th joint meeting of the CIDOC CRM SIG and SO/TC46/SC4/WG9; 38th FRBR – CIDOC CRM Harmonization meeting; the sig following the decision on issues 241 & 410 about merging the sections “Monotonicity”, “Minimality”, “Extensions”, “Coverage” and “Conservative Extension of the Scope of CIDOC CRM by Model Extensions” (modelling principles)in order to be included in the new version assigned HW to Steve to put them in an order, to PR and CEO to merge the text and MD to review the final text.
Heraklion, October 2019
posted by Steve on 9/12/2019
This has been sat on my desk for a couple of weeks and finally I have got round to typing up my thoughts.
I think it should be 2 sections: One covering Monotonicity and one covering all the different facets of Extensions. In the attached document I have coloured coded the current paragraphs into these two topics. Blue is extensions and green is monotonicity.
Hope that this helps
Posted by Martin on 9/12/2019
Sounds very good! I'll continue on that.
Posted by Martin 3/1/2020
I wish you all a Happy New Year, successful and in good health.
Here my attempt to describe the reality concept of the CRM and its relation to a knowledge base. Please comment!!
Reality and Knowledge Bases
The CIDOC CRM is a formal ontology in the sense introduced by N.Guarino [XXXX]. In order to understand the function of a formal ontology for collecting information in research processes about the past that can be shared, connected and integrated into coherent resources, one needs to make the following distinctions:
a) The material reality. For the purpose of the CRM, it is taken as that which is of a substance that can be perceived with senses or instruments, such as people, a forest or a settlement environment, sea, atmosphere, distant celestial or cellular micro structures, including what we assume that could be potentially or theoretically perceived if we could be there, such as the center of Earth or the sun, and all that is past. It is constraint to space and time. What is going on in our minds and produced by our minds is also regarded as part of the material reality, as it becomes materially evident to other people at least by our utterances, behavior and products.
b) The units of description or particulars, i.e., the things and relations as which we distinguish parts of reality when we refer to it, such as Mount Ida, the Taj Mahal, the formation of China by emperor Qin Shi Huang (秦始皇) in 221BC,Tut-Ankh Amun and his embalmment, Prince Shotoku of Japan sending a mission to China in 607AD, the participation of Socrates in the Battle of Potidaea or the radiocarbon dating of the Iceman Ötzi.
A formal ontology, such as the CIDOC CRM, constitutes a controlled language for talking about particulars. I.e., it provides definitions of classes and properties for categorizing particulars as so-called “instances” in a way that their individuation, unity and relevant properties are as unambiguous as possible. For instance, Tut-Ankh Amun as instance of E21 Person is the real pharaoh from his birth to death, and not extending to his mummy, as follows from the specification of the class E21 Person and its properties in the CRM.
For clarification, the CRM does not take a position against or in favor of the existence of spiritual substance nor of substance not accessible by either senses or instruments, nor does it suggest a materialistic philosophy. However, for practical reasons, it relies on the priority of integrating information based on material evidence available for whatever human experience. The CRM only commits to a unique material reality independent from the observer.
When we provide descriptions of particulars, we need to refer to them by unique names, titles or constructed identifiers, instances of E41 Appellation in the CRM, if the reference should be independent from context, such as reference by pronouns or enumerations of characteristic properties. The appellation itself, and the relation between the appellation and the referred item or relationship, must not be confused with the referred and its identity. Instances of the CRM are the real particulars, not their names. Particulars are approximate individuations, like sections, of parts of reality.
In contrast, a CRM-compatible knowledge base is an information object, instance of E73 Information Object in the CRM. It relates appellations with identifiers of CRM-Concepts in propositions about a described reality. Thereby users, in their capacity of having real-world knowledge and cognition, may be able to relate these propositions to the reality they are meant to characterize, and reason and research about their validity. In other words, the formal instances in a knowledge base are the identifiers, not the real things or phenomena. Therefore, a knowledge base does not contain knowledge, but represents knowledge of its maintainers, as long as there exist people that can resolve the used identifiers to their referents.
posted by Martin on 25/1/2020
I think the term "knowledge creation process" should be taken out of the "terminology" and be inserted in the Principles after "knowledge bases and Reality".
Please comment. If you agree, someone to review coherence with the other texts in this section please.
posted by Martin on 26/1/2020
Thank you very much. I have attended to your comments and one by George in the attached.
I accepted all changes, and put any change from the version 1_23_2020 in yellow, in order to make a new start.
I had thought about adding that Symbolic Objects may reside in the KB, but I think it makes the head of the reader explode. Should be introduced better in the scope note of Symbolic Object. I put the text in violet.
Thanasi, could you write up the texts for diagram 3 and 4? Could be quite similar to the previous texts. If you finish this, I could write up the STV stuff...
Please all to refer to this version _1_26_2020 (or the one from 1-23_2020 if you have made changes in there) from now on.
Posted by George on 26/1/2020
Have to say that the document is also confusing me at this moment. So I ended up writing texts for diagrams 1-3 in addition to the descriptionsn provided by Thanasis. Did not realize they were already written, as there was an open request to write them.
posted by Martin on 26/1/2020
No, I was confusing things. I suggest you, George, put on google docs the section from "We explain these concepts with the help of graphical representations in the next sections" on until the end, finish your version, and then others take up.
Posted by Thanasis on 4/2/2020
Here is my attempt to integrate the "knowledge creation process" entry into "knowledge bases". This is using Pat's latest version of the document.
Posted by CEO on 4/2/2020
I am sorry for not participating earlier. I have read the current final version of the document. It may be not always easy to understand even for me. Maybe some parts of the introduction to the "Principles for Modelling Ontologies: A Short Reference Guide could" could be used?
I will return with a more detailed suggestion tomorow.
Posted by Martin on 4/2/2020
I think we need to discuss this in the Meeting.
I think there are at least three to five different things we discuss here in a tangled way:
a) difference between understanding the laws that govern this world = the ontology, and the KB, which is a derivative of "I know this or this". We need to take a more clear position to what a justified believe means.
b) Who is the default agent expressing his knowledge in a KB.
c) How do we treat the possibility that any statement in a KB may turn out to be wrong, and by what sort of experience or process may that be revealed.
d) What is the role of formal reasoning on a system of statements that may individually be wrong, but in a majority expected to be true. Which deviations do we expect from an approximation in terms of some form of logic, and how this is usefully handled in practice.
f) Is there another logic, which would be superior to solving this issue, as Christian Emil points to below? I would not expect that. It sound like replacing true uncertainty with the certainty of an uncertainty model, for which we cannot find the parameters, because of uncertainty. I would rather expect that other forms of reasoning should be used. I would rather expect that some tolerance to some kinds of "inconsistencies" can be quite effective.
Posted by Martin on 6/2/2020
Attached my attempt to bring all bits and comments together in one consolidated version. I have reordered sections according to Pat's and Thanasi's comments.
I have rewritten Carlo's statement about contradictory KBs and reduced to cardinality violations.
I have inserted George's draft about spatial and temporal relations.
I kindly ask Thanasi to check if he would change these two sections.
I have not inserted George's draft about fig 1, but left my own version. I kindly ask George to check, if he would prefer his draft or improve my draft.
I will continue writing the section about spatiotemporal relations.
I kindly ask Nicola to create the promised RDF about Winkelmann, the full story as I refer to it, or just figure 2.
Since we cannot afford during the meeting making minor linguistic optimisations for such a long text, I ask in particular Steve to take over and do that before the meeting.
This text is mission critical for the next official release.
Please let me know ASAP, who can start these final revisions now. Best upload again to Google docs.
Further I attach a graphic designer's inspiration of the Winkelmann time-line (he died in Italy..). Please comment. I'd like to add that to the text, in some further improved version.
Posted by Christian Emil on 6/2/2020
I am sorry that my days are too busy due to practical application of CIDOC-CRM in archaeology. I mentioned that in my view parts of the document is not easy to understand. I you want to finish the work on the document, then my general comment is that we should not focus so much Knowledge bases. These are according to a widespread definition is a database/set/structure equipped with an interface/mechanism to to reasoning/interference and stem from Knowledge engineering and PROLOG in the 1980s. Knowledge bases are ok to argue the well definedness of the modeling principles to computer scientists and the like. But the topic is hard to understand for many museum persons and people doing cultural heritage documentation in general
Posted Thanasis on by 9/2/2020
Please find attached the text with some comments on George's sections
which apart from some editorial changes I think are good summaries of
how the various properties are meant to be used. I think the text was
cropped in the end. Happy to look at the remaining text if this is the case.
Posted by Martin on 9/2/2020
Thank you very much! The text is as I sent it. The section about spacetime is missing, I will write it now. I'll take over...
Posted by CEO on 10/2/2020
I have read the document (the latest version). It is in general clear. The part "Introduction to the basic concepts" is ok. We should add a figure 1b with events and actors.
I have some comments to the text in the first part. I have added comments not edited the text as the rule.
Posted by Carlo on 10/2/2020
I do not want to sound pedant or obnoxious, but I must say that the very first sentence ("The CIDOC CRM is a formal ontology in the sense introduced by N.Guarino”) gives me some problems. Nicola Guarino says that a formal ontology is a logical theory, and a logical theory is composed by a language (syntax and semantics) and a set of axioms. In the CRM specs, axioms are given using a special notation, but the language is definitely missing, to my understanding. Not to mention the fact that according to the nomenklatura, ontological axioms should be given in a modal logic explicitly formalising the notion of necessity. For instance, the last version of DOLCE that I read uses the language of S4 plus the Barcan formula.
I am a true friend of the CRM, deeply indebted to Martin and to the SIG for the amazing work you have been doing in the last twenty years. Yet, unless it is said somewhere and I missed it, there should be a sentence explaining why the CRM is not given as a set of axioms in a modal logic.
Posted by Christian Emil on 10/2/2020
According to OED obnoxious is defined as "Offensive, objectionable, odious, highly disagreeable. Now esp. (of a person): giving offence, acting objectionably; extremely unpleasant, highly dislikable. (Now the usual sense.)"
The easiest solution is to change this into into a reference t o the more computer science defintion of Gruber
"An ontology is an explicit specification of a conceptualization. The term is borrowed from philosophy, where an ontology is a systematic account of Existence. For AI systems, what “exists” is that which can be represented. (Gruber 1993, 1)"
and then write something like ' The CIDOC CRM can also be expressed/formalized as a theory in FOL, see ....
Posted by Martin on 10/2/2020
Thank you for pointing me to that. I usually refer to Nicola's paper about ontologies and information systems 1998.
I am a bit at a loss, why the CRM, once we have formulated it in FOL, and you have declared all the things about constants etc., does not have a language. Could you be more specific?
I assumed composing Class and property instances following the CRM axioms constitutes a language. Why not?
In my private communications with Nicola he pointed out that he does not care about encoding.
The question about modal logic is interesting. I think in terms of "possible states of affairs".
So, what should we say?:-)
Posted by Martin on 10/2/2020
There are arguments by Nicola, I would follow, that Gruber's is an underdetermined statement. I definitely like the distinctions Nicola makes in his paper from 1998. I miss the ontological commitment in Gruber's definition. That makes a big difference for me.
I'd rather prefer to be imprecise with "language", rather than talking about "conceptualizations" in an indeterminate way...
I'd rather be more precise about "ontological commitment" to an empirically understood reality, which can falsify, by experiment, a conceptualization.
Posted by Christian Emil on 10/2/2020
It is clear that Gruber's defintion is pretty vague. However, it has been a discussion among SIG members what is the authoritative definition of CRM. Most seem to consider the textual document to be the "real" defintion. The FOL formulation is considered by these to be an interpretation of the document as a FOL theory (or modal logic if that is preferred).
The FOL theory constitutes the formal framework without much semantics and is supplied with the rules of inference in FOL. The interpretation of the (valid) statements in this theory is described in the textual interpretation of the document. Correspondingly, more than one ontology can share the formal object oriented ISA hierarchy and properties between the classes. The semantics lies in the scope notes which describe our (the sig's) conceptualization of a part of the world.
My question is: Do we want the CRM to be the FOL interpretation or should it be the semi formal conceptualization described in the document?
Posted by Martin on 10/2/2020
I will probably never understand the following:
We define the CRM internally in TELOS. Then we create the document, replacing TELOS syntactic sugar with the known format in the document, which is a 1 to 1 mapping, another syntactic sugar. We do this for all releases.
Does this make the CRM semiformal, changing syntactic sugar, or is TELOS semiformal?
Is TELOS not a language? I always thought it is. If it is a language, why is my syntactic equivalent not a language?
Even if FOL has "few semantics", is there a measure how much semantics is needed to become "formal"?
How do you interpret "ontological commitment"???
How can any ontology be exclusively logic without an aboutness about the world?
I understand that the aboutness about the world, is an additional element to the logical form, and not a semiformalism.
Nicola Guarino describes in a naive way the real world to be a sort of LEGO set. That is nice, but he explicitly describes an aboutness about something which is not logic, but exists.
I perceive this world of reference as being non-discrete, but nevertheless endowed with a structure we approximate with the ontology.
Finally, how do you perceive a law of physics, such as Maxwell Equations. Is it formal, is it logical, is it about something that exists, that can be observed and verified experimentally, which must be of a nature that cannot be defined by formulae ("electric")?
Posted by Steve on 10/2/2020
In my opinion the CRM is "the semi-formal conceptualization described in
Posted by Carlo on 11/2/2020
> On 10 Feb 2020, at 19:11, Martin Doerr <email@example.com> wrote:
> Dear Carlo,
> Thank you for pointing me to that. I usually refer to Nicola's paper about ontologies and information systems 1998.
> I am a bit at a loss, why the CRM, once we have formulated it in FOL, and you have declared all the things about constants etc., does not have a language. Could you be more specific?
The paper does in fact provides an ontology in the more technical sense of the word (language plus axioms) although it does not use modal logic. In this sense, it complements the specs. This should be said. Please note that I am not doing this for having one more citation to the paper.
> I assumed composing Class and property instances following the CRM axioms constitutes a language. Why not?
It does, but the specs do not say what are the rules for “composing Class and property instances”. These rules are part of the grammar of the language. I assume these rules are very clear in your mind, however, they are not stated anywhere in the specs. In addition, there are the rules for stating the axioms, see below.
> In my private communications with Nicola he pointed out that he does not care about encoding.
But “encoding” means a lot. It may mean the specific concrete syntax one uses for, e.g. numbers or strings, in which case I agree it’s not so important in an ontology. But if it means the abstract syntax for expressing axioms, then it is of paramount importance. For example, by making the RDF abstract syntax a little bit too liberal, the W3C has created OWL Full, an undecidable monster that can express paradoxical classes such as the class of classes which are not instances of themselves. Of course the CRM specs do not need to revisit the history of logic, they just need to choose a decidable language that is enough expressive for the axioms it contains. See below.
> The question about modal logic is interesting. I think in terms of "possible states of affairs”.
Then you are intrinsically in the realm of possible world semantics, which does not sound surprising since everybody agrees with Carnap that meaning is strongly connected with necessity. You only need to make your thoughts explicit.
> So, what should we say?:-)
A quick synthesis of the paper would be my suggestion, it can be condensed in 3 lines. I can provide it if you want, I need some time to harmonise it with the rest of the section, but it has to wait until tomorrow afternoon.
Posted by Carlo on 11/2/2020
> On 10 Feb 2020, at 22:21, Martin Doerr <firstname.lastname@example.org> wrote:
> Dear All,
> I will probably never understand the following:
> We define the CRM internally in TELOS. Then we create the document, replacing TELOS syntactic sugar with the known format in the document, which is a 1 to 1 mapping, another syntactic sugar. We do this for all releases.
> Does this make the CRM semiformal, changing syntactic sugar, or is TELOS semiformal?
> Is TELOS not a language? I always thought it is. If it is a language, why is my syntactic equivalent not a language?
> Even if FOL has "few semantics", is there a measure how much semantics is needed to become "formal”?
TELOS is a language but only very few people know it. Logic on the other hand, is universally known, so if you want to communicate your ontology, you need to state it in a language that everybody understands. The alternative is that you include in the specs the TOIS paper that provides syntax and semantics of TELOS. This inclusion will settle the issue but I doubt that anyone would accept to learn TELOS in order to learn the CRM.
> How do you interpret "ontological commitment"???
Choosing a conceptualisation vs choosing another conceptualisation.
> How can any ontology be exclusively logic without an aboutness about the world?
I do not understand this question, let’s discuss it in person at the next SIG, I’m going to be there.
Posted by Christian Emil on 11/2/2020
I think we have a pedagogical issue here. Most peapol think that the TELOS is simply the format used in the definition tekxt of the CRM. We either don't know or have forgotten that there exist a reasoning mechanism in hte TELOS language/system.
Posted by Thanasis on 11/2/2020
As always apologies for the ignorance, I have been spotting references
to TELOS for many years and I always thought it was a piece of internal
software at FORTH for managing ontologies. Is there a reference for an
introduction to TELOS?
Posted by Martin on 11/2/2020
Dear Carlo, Christian-Emil,
Thank you very much for your comments:-). If you, Carlo, would add a few lines about the CRM being a formal ontology in the sense of Guarino's paper from 1998, I would appreciate. I do not remember Nicola's definition to require a specific logical language, syntax or modal logic.
We address with the CRM people that have no idea about these things, and I truely believe that with all the efforts to provide FOL definitions, also the logicians and IT guys can guess the missing details.
Can we be as unprecise as possible, without causing ambiguities of substance?
I would indeed expect that any logical theory established to formulate any ontology would allow for multiple ontologies to conform with the same logical formalism.
For me, of uttermost importance is the interpretation function between the "domain" and the ontology in Nicola's paper, and the fact that he states that the logical theory DOES NOT define the meaning completely, but that it is an approximation to reduce the residual unintended models.
My understanding is, that the scope notes CONSTITUTE the interpretation function. IF the domain is not a naive LEGO set, this interpretation function CANNOT be a logical theory. But this does not make the logical theory "semiformal".
For instance, I have not been convinced so far that describing the axioms of parthood, which you can find in many publications, allows anyone to detect parthood in the real world, and that there exists no other relation with the same axioms, but not being parthood. Parthood implies the axioms, and the axioms select a specific kind of parthood, but parthood cannot be inferred from the axioms, I believe.
Same with "E21 Person", etc...
This goes far beyond Thomas Gruber, and makes a difference to contructivism.
Ontological commitment for me has to do with the interpretation function, but we may look at that in more detail.
You say "Choosing a conceptualisation vs choosing another conceptualisation." What is the criterion? choosing for to do what?
Looking forward to discuss in person....
Posted by Martin on 11/2/2020
Attached. John Mylopoulos, Alexander Borgida, Matthias Jarke, Manolis Koubarakis:
Telos: Representing Knowledge About Information Systems. ACM Trans. Inf. Syst. 8(4): 325-362 (1990)
Another comprehensive review will appear this spring.
Telos. Encyclopedia of Database Systems (2nd ed.) 2018
Bryan M. Kramer, Vinay K. Chaudhri, Manolis Koubarakis, Thodoros Topaloglou, Huaiqing Wang, John Mylopoulos:
Implementing Telos. SIGART Bulletin 2(3): 77-83 (1991)
Posted by Christian Emil on 11/2/2020
To act as the Devil's Advocate: Most readers of the CRM don't care very much about the first order logic interpretations or interpretations in any other formal deduction systems. That is for the advanced and very skilled readers.
Most reader try to understand how the CRM can be used to publish open data so that they can be linked easily and efficient. in the same way people try to use RDF or graph databases as a handy way to make a common and flexible index for heterogeneous databases.
Said so, my view is that the definition of CRM should be formulated in such a way that it can accessible and understandable for people looking for an relatively easy and exhaustive integration tool, but it should also have a stringent logical definition which make it possible to implement it in an information system with deduction (also called Knowledge Base). So we should never make it a necessary requirement that the readers have skills in formal logic. If we do so our usergroup will not be very large. However, the skilled readers should see that the definitions also works on a formal logical level and can easily be implemented in formal systems.
Posted by Martin on 11/2/2020
I share exactly your point of view.
Posted by Christian - Emil on 11/2/2020
Excellent, then what remains is to edit the modeling principles chapter to make adequate for the two reader groups. This is pedagogical task and not so easy.
Posted by Carlo on 12/2/2020
In the meantime, I found Guarino’s 1998 paper, from which I cut the following:
In the philosophical sense, we may refer to an ontology as a particular system of categories accounting for a certain vision of the world. As such, this system does not depend on a particular language: Aristotle’s ontology is always the same, independently of the language used to describe it. On the other hand, in its most prevalent use in AI, an ontology refers to an engineering artifact, constituted by a specific vocabulary used to describe a certain reality, plus a set of explicit assumptions regarding the intended meaning of the vocabulary words. This set of assumptions has usually the form of a first-order logical theory, where vocabulary words appear as unary or binary predicate names, respectively called concepts and relations. In the simplest case, an ontology describes a hierarchy of concepts related by subsumption relationships; in more sophisticated cases, suitable axioms are added in order to express other relationships between concepts and to constrain their intended interpretation.
The two readings of “ontology” described above are indeed related each other, but in order to solve the terminological impasse we need to choose one of them, inventing a new name for the other: we shall adopt the AI reading, using the word conceptualization to refer to the philosophical reading. So two ontologies can be different in the vocabulary used (using English or Italian words, for instance) while sharing the same conceptualization.
The paper (in the revised form) is attached.
Posted by Christian-Emil on 12/2/2020
I have experienced that most of my colleagues in the humanities get lost at parenthesis level zero when it comes to logical expressions.
It may be easier to write separate texts for the logic geeks and the rest.
Posted by Carlo on 12/2/2020
I’ve expanded the first paragraph of the first section following our previous discussion. This addition, in my intentions, settles the “formal ontology” issue and makes any reference to or quotation of the logical formulae superfluous.
Posted by Carlo on 14/2/2020
BTW, I think I now understand Martin's objection on the fact that an ontology needs to have some aboutness about the world.
My answer is that logical axioms express all the aboutness a computer can understand. For humans, you can use scope notes, i.e., natural language, because humans share the world, have experiences in it, and so can (more or less) understand what you mean by an expression such as (I choose a class at random) "This class comprises all phenomena, such as the instances of E4 Periods, E5 Events and states, which happen over a limited extent in time”. We all know what a phenomenon is, what “happen” means and so on. But computers don’t because they do not have access to the world, so all we can tell them to explain what we mean are the logical axioms.
Posted by Martin on 14/2/2020
Before integrating Christian-Emil's and carlo's latest proposals, here my editing of spatial, temporal and writing spatiotemporal relations.
Please have a look here, particularly the last one. Exhausted... We have forgotten P170 defines time....
Posted by Martin on 14/2/2020
On 2/14/2020 7:30 PM, Carlo Meghini at ISTI CNR wrote:
> BTW, I think I now understand Martin's objection on the fact that an ontology needs to have some aboutness about the world.
> My answer is that logical axioms express all the aboutness a computer can understand. For humans, you can use scope notes, i.e., natural language, because humans share the world, have experiences in it, and so can (more or less) understand what you mean by an expression such as (I choose a class at random) "This class comprises all phenomena, such as the instances of E4 Periods, E5 Events and states, which happen over a limited extent in time”. We all know what a phenomenon is, what “happen” means and so on. But computers don’t because they do not have access to the world, so all we can tell them to explain what we mean are the logical axioms.
Exactly, but data entry is done by humans:-) for humans....
So, first come people, logic assists...logic needs to conform with experiment. At least in science. If not, change logic until it does.
Posted by Martin on 15/2/2020
Here my attempt to integrate everything, cardinalities corrected, P170 introduced to temporal...
Chrysoula, please use this for the meeting.
Posted by Carlo on 17/2/2020
> Exactly, but data entry is done by humans:-) for humans….
So we need a human-readable/usable version, that’s clear.
> So, first come people, logic assists...logic needs to conform with experiment. At least in science. If not, change logic until it does.
I’d say that the logic we have today reflect the work of so many scientists and philosophers that I would think twice before changing it, and after thinking twice, I’d leave it as it is. I speak for myself of course
I am not sure you got my proposed modification of the first sentence, so I’m re-sending it. I’ve deleted from the attached document all the rest of the chapter to make this email lighter.
Posted by Martin on 17/2/2020
On 2/17/2020 11:52 AM, Carlo Meghini at ISTI CNR wrote:
>> Exactly, but data entry is done by humans:-) for humans….
> So we need a human-readable/usable version, that’s clear.
Not only, the human user can understand more instructions about reality, and hence be more precise about the intended reality. So I perceive the scope notes as the additional instructions for the humans, to enter correct data and the understand query results correctly.
In general, they will be able to understand the deviations the applications of logic create from the intended reality. For instance, we are used to query Open World data with closed world SPARQL, even though it has no concept for absence of knowledge.
>> So, first come people, logic assists...logic needs to conform with experiment. At least in science. If not, change logic until it does.
> I’d say that the logic we have today reflect the work of so many scientists and philosophers that I would think twice before changing it, and after thinking twice, I’d leave it as it is. I speak for myself of course ;)
I wanted to say: Use another kind of logic of mathematical formalism for the problem. Not "change logic itself". I mean things like fuzzy logic, second order, probabilistic models, partial differential equations, integral equations, complex functions, algebra of groups etc. Physicists invented some decades ago a theory of functions f(x) which are zero at all points x except for x=0, where they are infinite, and integral -inf to +inf = 1. There was no other way. Mathematicians first resisted, then accepted, and then formalized. Algebra of groups was invented by mathematicians, and physicists found them useful in quantum mechanics. So what? The problem just dictated another formalism. I have used in the past all these except for FOL and fuzzy logic.
> I am not sure you got my proposed modification of the first sentence, so I’m re-sending it. I’ve deleted from the attached document all the rest of the chapter to make this email lighter.
It's in the final I sent around
Posted by Carlo on 17/2/2020
Thank you Martin, I think we agree. Once someone made a proposal, to “logicize” the scope notes. I think it’s a valid proposal! by doing this we will discover the language the SIG uses to tell the meaning of CRM to the users, and a whole new ontology
In the 46th joint meeting of the CIDOC CRM SIG and ISO/TC46/SC4/WG9; 39th FRBR - CIDOC CRM Harmonization meeting; The sig went through the text of the introduction to the CRM section that has to do with the modelling principles and the examples supplied and reviewed the changes that the reviewers proposed. The sig only discussed parts of the text that had previously been judged unclear by the reviewing team. The text in its final form appears here.
Overall decision regarding the introductory section of the CRM: Given that the document still needs editing it won’t be ready to be submitted at ISO before the 47th sig meeting. In view of that, it was decided that the next version (7.0) will become the official version for the sig community (because we can’t rely on 6.2.1 with all the changes implemented over the last year). This version will appear in RDF as well. Extensive email votes will be held in the period leading up to the next meeting.
Following the next sig meeting, updates in the introduction will also be incorporated in the official version (v.7.1) which will then be submitted to ISO.
Other additional decisions are:
-1. Reality, Knowledge Bases and CIDOC CRM (first paragraph) : The sig appointed MD to rephrase the first paragraph (see below) of this section [HW]
a. The reformulation should not refer to a “vocabulary” –to avoid confusion with thesauri (even though this is an unwarranted connection). Also, the reference to the definition of formal ontology by Guarino should be made into a footnote.
b. It should be made clear what is meant by “non-technical audience”, namely they are not necessarily computer scientists.
- provide links or references to the classes and properties’ definitions mentioned throughout the introduction, given that they have not been previously introduced in this section of the document
- The title of the section "Knowledge Creation Process of Knowledge Base Contents" has been renamed “Authorship of Knowledge Base Contents”.
- The following sentence needs be further revised; it must be made clear that incompleteness of information is not the same as the existence of contradictory statements within a DB. CEO is on the rewrite [HW]. MD proposed that it be made into a footnote. Once it’s been edited, it will be put up for an e-mail vote.
"Statements in a KB may be in contradiction to the ontologically defined quantification of properties without the KB being broken or invalid in any sense, either because necessary properties are unknown or there exist good reasons to assume alternative values for properties with limited cardinality, be it by the same or by different maintainers"
- hierarchical relations of specialization are to be constantly abbreviated by IsA throughout the text.
- Introduction to the basic concepts,Figure 2: Winkelmann sees the statue of Laocoon: The sig agreed with the diagram provided and the text explaining it . Minor suggested editions involve clarifying that the thumbnail pictures represent the person/object and do not stand for the visual items representing said entities (maybe in a footnote).
- Introduction to the basic concepts,Figure 3: Figure depicting how the paths of Winkelmann and the statue of Laocoon cross in space and time (i.e. how the spacetime volumes of Winkelmann and Laocoon’s statue meet). The figure is going to be incorporated in the CIDOC CRM definition (introduction to the basic concepts section). However, the specifics of the representation need be reworked, because the diagram in its current form gives the wrong impression regarding the direction of movements and the succession of movements from one place to the next. HW: ML is to contact KD to provide a better visualization for that.
- Figure 3 (current): Spatial information: The sig reviewed the diagram and accepted it as is. It will be renamed “Figure 4: reasoning about spatial information”.
- Figure 4 (current) will be renamed “Figure 5 reasoning about temporal information”
- Figure 5 (current) will be renamed “Figure 6 reasoning with spacetime volumes”
The minutes of this discussion can be found here.
Athens, February 2020
Posted by Martin on 16/3/2020
Please update 7.0 with the attached slide 2,4. Changes: style of multiple properties, P167 added to slide 2, following meeting decision.
Posted by Martin on 30/3/2020
Dear All attached my attempt to resolve (see minutes):
1. Reality, Knowledge Bases and CIDOC CRM (first paragraph)
Posted by Thanasis on 6/4/2020
I think the footnoted version works. I would just replace "embrace" with "for".
Posted by Christian Emil on 29/4/2020
As a part of my homework I should comment on Martin's three alternatives for the first paragraph of the Introduction text see
Martin's alternatives can be found at http://cidoc-crm.org/sites/default/files/459%20Guarino_Definition%20HW%20MD%202.docx
The current working document is verion 6.2.8 (on cidoc-crm.org it is 6.2.7) so I copy in the text from the 6.2.8
The following modelling principles have guided and informed the development of the CIDOC CRM.
Reality, Knowledge Bases and CIDOC CRM
The CIDOC CRM is a formal ontology in the sense introduced by N.Guarino. that is a specific vocabulary used to describe a certain reality, plus a first-order logical theory narrowing down the intended meaning of the vocabulary words (N.Guarino 1998). The syntax and formal semantics of this first-order theory are given in (Meghini & Doerr 2018), where the computational aspects are also discussed. The present document is intended for a non-technical audience, therefore it focuses on the informal semantics and on the pragmatics of the CIDOC CRM vocabulary, offering a detailed discussion of the main traits of the conceptualization underlying the CIDOC CRM through the basic usage patterns.The CIDOC CRM aims to assist sharing, connecting and integrating information from research about the past. In order to understand the function of a formal ontology of this kind, one needs to make the following distinctions :"
My view, which I also have the impression that is the common view, is that text should be as easy as possible. I therefore suggest that the foot note version (see the link and the text I copied in at the end of this email) should be used:
The CIDOC CRM is a formal ontology in the sense introduced by (N. Guarino 1998). The present document is intended for an audience not specialized in computer science and logic; therefore, it focuses on the informal semantics and on the pragmatics of the CIDOC CRM concepts, offering a detailed discussion of the main traits of the conceptualization underlying the CIDOC CRM through the basic usage patterns. The CIDOC CRM aims to assist sharing, connecting and integrating information from research about the past. In order to understand the function of a formal ontology of this kind, one needs to make the following distinctions :"
 Nicola Guarino defines a formal ontology as a specification of a set of named concept used to describe and approximate a part of reality, plus a first-order logical theory narrowing down the intended meaning of the named concepts.
 For the readers interested in computer science and logic, the syntax and formal semantics employed by the CIDOC CRM are given in (Meghini & Doerr 2018), where the computational aspects are also discussed.
In the virtual meeting of editorial group of CIDOC-CRM version 7.0 on 11/5/2020, the editorial group reviewed Martin's Hw on 30/3/2020 and changed the proposed text to
"The CIDOC CRM is a formal ontology in the sense introduced by (N. Guarino 1998) [fN 1]. The present document is intended to embrace an audience not specialized in computer science and logic; therefore, it focuses on the informal semantics and on the pragmatics of the CIDOC CRM concepts, offering a detailed discussion of the main traits of the conceptualization underlying the CIDOC CRM through basic usage patterns [fN 2]. The CIDOC CRM aims to assist sharing, connecting and integrating information from research about the past. In order to understand the function of a formal ontology of this kind, one needs to make the following distinctions:
[fN 1]: Nicola Guarino defines a formal ontology as a specification of a set of named concepts used to describe and approximate a part of reality, plus a first-order logical theory narrowing down the intended meaning of the named concepts.
[fN 2]: For the readers interested in computer science and logic, the syntax and formal semantics employed by the CIDOC CRM are given in (Meghini & Doerr 2018), where the computational aspects are also discussed. "
Editorial Group of CRMbase v.7.0
Sent by Martin to editorial group of CIDOC-CRM version 7.0 on 13/5/2020
<Introduction to the basic concepts, in the middle of 6th paragraph>
Here my attempt to be more explicit:
Identity in the sense of the CIDOC CRM, therefore, means that informed people are able to agree that they refer to the same, single thing, that according to the scope note of the respective CIDOC CRM class it is regarded to be an instance of.
Identity in the sense of the CIDOC CRM, therefore, means that informed people are able to agree that they refer to the same, single thing in its distinction from others, its extent and over its time of existence. The criteria for such a determination should come from understanding the scope note of the respective CIDOC CRM class this thing is regarded to be an instance of, because communication via information systems may not leave space for respective clarifying dialogues between users.
Is this better?
Posted by Christian Emil on 21/6/2020
Please find a document with the HW for the issues 456 and 459. These issues are about changes in the text of the introduction to the CRM. The document will be used in the virtual SIG meeting