Dear Richard, On 12/21/2017 5:38 PM, Richard Light wrote:
I actually did not propose to add new classes and properties. Identity conditions are rules that are not explicit constructs in the CRM so far, but in scope notes. In general, all class scope notes should describe identity conditions. That is an issue for the next versions.Martin,I'm not sure I accept the idea that strings have an identity, which is separate from the way in which they happen to have been used, and which it is worth our while to record by adding new classes and properties to the CRM.
Right. That is my point. Clearly, there is an identity condition of a string, which is only the sequence of symbols and the relevant encoding level. For instance, is an ASCII "H" the same with a Latin 1 "H" and - or not - with a Greek capital "H" ? Obviously there are different choices, making sense depending on what we want to describe.Any string value which is of interest to us will be identified as an instance of a suitable CRM class: E35 Title; E41 Appellation; etc. Staying with titles for now, if we know who assigned a title to an art work, we have E13 Attribute assignment which allows us to state what we know (who assigned the title, when, etc.); otherwise we can put a simple P1 is identified by property to join the art work itself directly to its title. Either way, the title now has a context (i.e. a number of CRM assertions which reflect what we know about the art works bearing that title). In this particular case, what more is there to say?Another example of this sort of thing (which might be more convincing) would be surnames. There are lots of resources which detail the origin, meaning and geographical spread of peoples' family names. Each name is a subject of study in its own right. Looking at the definition of E41 Appellation, I notice that it relates to "a specific instance of some class or category within a certain context". That isn't quite the same as the surname /as a name/.
If we talk about names, is William and Bill the same? Obviously there are different choices, making sense depending on what we want to describe.
If we are on the level of titles, is the identity condition only that of a string, or of a string and...
If we are in Information Objects, is the identity condition only that of a string, or of a string and...?
The problem occurs only with short strings with a good chance to be the same at some encoding level, but not in intended meaning. It may be marginal. I don't know. But translations of titles clearly use more identity conditions than the string.
All the best, Martin
Richard On 20/12/2017 17:22, Martin Doerr wrote:Dear Richard, On 12/20/2017 6:23 PM, Øyvind Eide wrote:I think we are confusing terms here. RDF propositions are "context free" because they use URIs, which should(!) resolve unambiguously to the intended reference, without needing a context of utterance to grasp the meaning. Clearly, information objects are in general not context free, how could they! The question I raised is, which parameters of the context of either use or creation are the ones in our cultural communication practice, which allow us to differentiate between two identical strings with obviously two different intended meanings. This context is well acknowledged in the information theory by Shannon, who simply assumes an a priori unambiguous agreement between sender and receiver about the employed signs. If we have understood which context parameters are relevant, and which cases to distinguish, we will formulate them in the form of context-free propositions. Normally, we associate information objects with the metadata about their creation, which is one form of specifying a context. However, creation may be unknown, as in the case of fairy tales, and yet they have an identity. So, it's more complicated.Am 19.12.2017 um 12:10 schrieb Richard Light <rich...@light.demon.co.uk <mailto:rich...@light.demon.co.uk>>:Introducing a class for context sounds strange to me as it would indicate that the rest is context-free. Still, it could make sense in order to have a possibility to make context explicit in the cases we need to.All the best, martinSurely the statements which we make with the CRM are themselves the 'context' for individual assertions? If so, we have our context already, and don't need to invent an artificial mechanism to express it.I would say ONE context, not THE context. The assertions often come from somewhere, loosing one context and adding a new one, the two being more or less similar.Granted, the original context is often lost already when the data was entered into a database, long before it was expressed in CRM.A series of de/re-contextualisations… Regards, ØyvindRichard On 18/12/2017 10:05, Øyvind Eide wrote:Am 15.12.2017 um 10:53 schrieb Martin Doerr <mar...@ics.forth.gr <mailto:mar...@ics.forth.gr>>:On 3/26/2017 9:29 PM, Øyvind Eide wrote:Dear Martin, this is dangerous territory. Do we need to go there? We may have to open up all sorts of boxes including those owned by language philosophers and semioticians. An utterance is made by someone, surely. But is a title an utterance? It is not purely either or, but is it not more langue than parole?https://en.wikipedia.org/wiki/Langue_and_parole I think one can find many different views on what information is in the humanities and many of them would be quite different from Shannon. Personally, I think thinking based on dialogism makes a lot of sense. Do we have to enter this territory? Do we need to express opinions on these things in CRM?Dear Øyvind,Clearly, one principle of the CRM is, never interpret a term! So, we are not concerned settling disputes about what information or an utterance is. We are concerned with the consistency and effectiveness of definitions for our information purposes. So, for me the problem is a simple question of disambiguation of identity.Since you wrote (and I agree) "E35 Title can only be used when such a string is actually are used as a title...." this implies that (a) the same string may be used twice as a title and (b) translates differently in these cases.This means, that the identity of the title as described above consists of the string + context. Otherwise, the scope note is inconsistent. This context can either be determined as (1) language, (2) one work of art, (3) multiple works of art intentionally referring to the same source - F1 Work or"loans" from other F1 Work.This creates a precedent with respect to identity of information. Equally obviously, if we create in the CRM an identifier for "mehr Licht" by Goethe, true or not, and want to trace arguments about the interpretation and reality in an information system, we must, if we want or not, carry the context with us. So, we have two choices: Either we keep the identity of an E73 provenance independent, and introduce another class for information object use context, or we imply a concept of provenance as part of the identity of the information object.Equally obviously, it is impossible in general to trace exact provenance. We could, however, in the scope note, describe the context concept behind an information object in a more general way, which implies specialization from case to case.A relevant application are tombstone and other short inscriptions. Epigraphy experts regard the same text on another stone as different.We may even talk about two message levels. For instance "r.i.p." as a generic message in the tombstone context, and "r.i.p." as a personal message on aparticular tombstone. Or we say r.i.p. to the issue;-)Indeed, Martin. I see your arguments, and hopefully understand them in the right context.As Hirst pointed out, context is a spurious concept. We need some, we never need (or can have) all, and the border between the two is unsharp.Is this not also the trade-off of information integration in general, and where we disagree with the semantic web community (a sentence that should have had a lot of qualification)? Because we know that the dream of a decontextualised emerging network of useful information is just a dream, at least for cultural history and the humanities. Still, we also know that if we let ourselves tie down to traditional levels of context we are lost and will never be able to integrate something.Introducing a class for context sounds strange to me as it would indicate that the rest is context-free. Still, it could make sense in order to have a possibility to make context explicit in the cases we need to.Regards, ØyvindCheers, martinRegards, Øyvind24. mar. 2017 kl. 12.50 skrev martin<mar...@ics.forth.gr>: Dear Oeyvind, I agree with the scope note, given the interpretation we decided. I wonder however if there is a deeper issue here: In Germany there exists the saying that dying Goethe uttered "mehr Licht" ("more light"). I reused this proposition yesterday, because I wanted to read a newspaper. Claude Shannon defined information as a message with a known provenance, which is the most accepted theory in computer science. That would mean that the identity of an Information Object is a tuple <content,sender>, rather than <content>. If we accept that, we enter another hell of arguments about what the identity of the sender is. That is easy for a Title, but quite tricky for the non-smoking symbol. Question: Should we touch also this front, or are we sure that "more light" is always "more light" ? In other words, may be a title actually deviates from an appellation in that it adds to its identity the provenance, which in turn allows for translation? best, martin On 24/3/2017 11:45 πμ, Øyvind Eide wrote:Dear all, Here is my homework for Issue 260: 1. E35: Accepted the comment made by Oyvind that the scope note of E35 Title is misleading, since it refers to something functioning a title, not having the form of a title, it is decided to keep the Title, to update scope note. This HW is assigned to Oyvind I have changed the first paragraph of the scope note Old scope note for E35:This class comprises the names assigned to works, such as texts, artworks or pieces of music. Titles are proper noun phrases or verbal phrases, and should not be confused with generic object names such as “chair”, “painting” or “book” (the latter are common nouns that stand for instances of E55 Type). Titles may be assigned by the creator of the work itself, or by a social group. This class also comprises the translations of titles that are used as surrogates for the original titles in different social contexts.Proposed new version: “This class comprises textual strings that within a cultural context can be clearly identified as titles due to their form. Being a subclass of E41 Appellation, E35 Title can only be used when such a string is actually are used as a title of a work, such as a text, an artwork, or a piece of music. Titles are proper noun phrases or verbal phrases, and should not be confused with generic object names such as “chair”, “painting” or “book” (the latter are common nouns that stand for instances of E55 Type). Titles may be assigned by the creator of the work itself, or by a social group. This class also comprises the translations of titles that are used as surrogates for the original titles in different social contexts.” ————————— 2. E49 Time Appellation: to keep but it should be merged with Date and it should be decided if they keep the same name (Oyvind) E50 Date should be marked obsolete. I have changed the inheritance, the first paragraph of the scope note, and added two examples. Old definition of E49 Time Appellation:Subclass of : E41 Appellation Superclass of: E50 Date Scope Note: This class comprises all forms of names or codes, such as historical periods which are characteristically used to refer to a specific E52 Time-Span. This includes human- and machine readable dates and timestamps. The instances of E49 Time Appellation may vary in their degree of precision, and they may be relative to other time frames, “Before Christ” for example. Instances of E52 Time-Span are often defined by reference to a cultural period or an event e.g. ‘the duration of the Ming Dynasty’. Examples: • “Meiji” [Japanese term for a specific time-span] • “1st half of the XX century” • “Quaternary” • “1215 Hegira” [a date in the Islamic calendar] • “Last century”New definition of E49 Time Appellation: Subclass of : E41 Appellation Scope Note: This class comprises all forms of names or codes, such as historical periods, and dates, which are characteristically used to refer to a specific E52 Time-Span. The instances of E49 Time Appellation may vary in their degree of precision, and they may be relative to other time frames, “Before Christ” for example. Instances of E52 Time-Span are often defined by reference to a cultural period or an event e.g. ‘the duration of the Ming Dynasty’. Examples: • “Meiji” [Japanese term for a specific time-span] • “1st half of the XX century” • “Quaternary” • “1215 Hegira” [a date in the Islamic calendar] • “Last century” • “2013-10-05” • “Mon May 19 22:39:23 CET 2014” Kind regards, Øyvind _______________________________________________ Crm-sig mailing list Crm-sig@ics.forth.gr http://lists.ics.forth.gr/mailman/listinfo/crm-sig-- -------------------------------------------------------------- Dr. Martin Doerr | Vox:+30(2810)391625 | Research Director | Fax:+30(2810)391638 | | Email:mar...@ics.forth.gr | | Center for Cultural Informatics | Information Systems Laboratory | Institute of Computer Science | Foundation for Research and Technology - Hellas (FORTH) | | N.Plastira 100, Vassilika Vouton, | GR70013 Heraklion,Crete,Greece | | Web-site:http://www.ics.forth.gr/isl | -------------------------------------------------------------- _______________________________________________ Crm-sig mailing list Crm-sig@ics.forth.gr http://lists.ics.forth.gr/mailman/listinfo/crm-sig-- -------------------------------------------------------------- Dr. Martin Doerr | Vox:+30(2810)391625 | Research Director | Fax:+30(2810)391638 | | Email:mar...@ics.forth.gr | | Center for Cultural Informatics | Information Systems Laboratory | Institute of Computer Science | Foundation for Research and Technology - Hellas (FORTH) | | N.Plastira 100, Vassilika Vouton, | GR70013 Heraklion,Crete,Greece | | Web-site:http://www.ics.forth.gr/isl | --------------------------------------------------------------_______________________________________________ Crm-sig mailing list Crm-sig@ics.forth.gr http://lists.ics.forth.gr/mailman/listinfo/crm-sig-- *Richard Light* _______________________________________________ Crm-sig mailing list Crm-sig@ics.forth.gr <mailto:Crm-sig@ics.forth.gr> http://lists.ics.forth.gr/mailman/listinfo/crm-sig_______________________________________________ Crm-sig mailing list Crm-sig@ics.forth.gr http://lists.ics.forth.gr/mailman/listinfo/crm-sig-- -------------------------------------------------------------- Dr. Martin Doerr | Vox:+30(2810)391625 | Research Director | Fax:+30(2810)391638 | | Email:mar...@ics.forth.gr | | Center for Cultural Informatics | Information Systems Laboratory | Institute of Computer Science | Foundation for Research and Technology - Hellas (FORTH) | | N.Plastira 100, Vassilika Vouton, | GR70013 Heraklion,Crete,Greece | | Web-site:http://www.ics.forth.gr/isl | -------------------------------------------------------------- _______________________________________________ Crm-sig mailing list Crm-sig@ics.forth.gr http://lists.ics.forth.gr/mailman/listinfo/crm-sig-- *Richard Light* _______________________________________________ Crm-sig mailing list Crm-sig@ics.forth.gr http://lists.ics.forth.gr/mailman/listinfo/crm-sig
--
--------------------------------------------------------------
Dr. Martin Doerr | Vox:+30(2810)391625 |
Research Director | Fax:+30(2810)391638 |
| Email: mar...@ics.forth.gr |
|
Center for Cultural Informatics |
Information Systems Laboratory |
Institute of Computer Science |
Foundation for Research and Technology - Hellas (FORTH) |
|
N.Plastira 100, Vassilika Vouton, |
GR70013 Heraklion,Crete,Greece |
|
Web-site: http://www.ics.forth.gr/isl |
--------------------------------------------------------------
