I wanted to comment on email exchanges that have taken place over recent times. CIDOC CRM is a major achievement, not just in cultural heritage, but in the humanities generally. The founding members of the CRM and others that joined them in the CRM SIG achieved what Knowledge Representation (KR) computer scientists were unable to achieve when they first described KR. Their concept of KR was clear - it relies on making a connection between “problem solving expertise” and “the domain” in which this problem solving takes place. The key element was the latter, which needed to provide a solid framework of representing reality.
Without a reality based domain framework they knew that integration of information could not be achieved - they understood that it was impossible to effectively integrate on an epistemological basis (see Davis, Shrobe, Szolovits (1993), What Is a Knowledge Representation). However many computer scientists were unable to undertake this arduous task of describing the reality of a domain (and they explicitly replaced ‘ontology’ with ‘schema’ to clarify that they had removed this burden - although this naming practice has not been continued) and focused on reasoning experiments with epistemological information. The CRM SIG achieved what computer scientists couldn’t themselves, and didn’t, because it required a significant empirical investigation demanding many years of work in another domain. The CIDOC CRM attempts to find points of integration across and within different disciplines. These points of integration are the connection between reality and practice. CIDOC CRM concepts need to be both universally accepted and be capable of being universally applied. This is what provides a robust semantic framework for integrating information. In the highly fragmented, disjointed and unreliable world of the digital and the Web, the CRM stands out as a robust framework for interdisciplinary work and engagement. We know only too well in cultural heritage that the efforts to aggregate the different structures and languages with different epistemological traditions without an ontological framework has only led to reductive implementations that are of limited value long term. The construction of our integrating ontology, which is one of relatively few that complies with the original KR blueprint, could only have been achieved with the huge effort of CRM SIG members and in particular those founding members who established the fundamental key principles. We should consider the increased activity in developing more specialisations as an indication of the success of the CIDOC CRM - but these specialisations cannot be epistemological constructions regardless of how probable or possible they are. They must be compatible with the principle of domain reality through robust empirical examples and practice. Without it, the ontology just becomes yet another schema, and will become fragmented like the academy. The CRM SIG works on clear principles and not on the invention of schemas to satisfy pragmatic individual needs, meet artificial project deadlines, support pragmatic epistemology bubbles, or satisfy commercial contracts. The CRM is based on transparency and understanding, not closed shops. In the midst of this a worrying trend has appeared in that the CRM SIG has become, in some instances, an environment which prevents open discussion and which lacks the appropriate professional standards and courtesy needed in communication, and this is not appropriate. The principles and methodology of the CIDOC CRM are clear and have been described in much detail over many years. The founding designers of the CIDOC CRM are people who have wide interests, discuss many different subjects and are very open and accessible. For a recent example, I discussed with a CRM colleague the subject of social class, which is of great importance to me and historians tracking social change and transition. I have very clear thinking about, and use the concept of, social class, and other aspects of society which are an important part of my work, but I doubt that everyone would agree with my conceptions. We can talk and agree about their importance, and existence, but they are not concepts that can be defined under the CIDOC CRM methodology and framework which require clear boundaries. For example, from Bourdieu’s essay “What Makes a Social Class? On The Theoretical and Practical Existence Of Groups”, “In the reality of the social world, there are no more clear boundaries, no more absolute breaks, than there are in the physical world. The boundaries between theoretical classes which scientific investigation allows us to construct on the basis of a plurality of criteria are similar, to use a metaphor of Rapoport's, to the boundaries of a cloud or a forest. These boundaries can thus be conceived of as lines or as imaginary planes, such that the density (of the trees or of the water vapour) is higher on the one side and lower on the other, or above a certain value on the one side and below it on the other. (In fact, a more appropriate image would be that of a flame whose edges are in constant movement, oscillating around a line or surface.) Now, the construction of (mobilized or "mobilizable") groups, that is, the institutionalization of a permanent organization capable of representing them, tends to induce durable and recognized divisions which, in the extreme case, i.e., at the highest degree of objectification and institutionalization, could take on the form of legal frontiers." To repeat, this doesn't mean that they are not important (they are hugely important) or not valid or that they cannot be represented, or that they should not be talked about. They can be appropriately modeled (and their status made clear) and they can be referred to by the CIDOC CRM with a requisite interface, as we do for terminology (E55 Type). These interfaces are something we can work on. But they are not compatible with the principles of the CIDOC CRM and therefore cannot be specialisations of the base CRM. The CIDOC CRM does not deal in “a probable class” or the “probability of a real class” and that is its limitation - but also its strength in that it provides a framework for serious robust contextual data integration missing from traditional CH databases. On this point the narrative recently provided on CRMSoc which talks about socially constructed facts seems to fall into this incompatibility category. That doesn’t make it invalid or not a valuable contribution, but, depending on the definition, it is likely to make it incompatible (unless, for example, legal identities are created) as a CIDOC CRM specialisation proposal because it appears to take a position that cannot be adequately evidenced or accepted across the domain, or even within one discipline (for example History), particularly because of its postmodern associations. Social constructivism itself has no single agreed definition and has relativist elements (for example see, The Good, the Bad, and the Ugly: The Many Faces of Constructivism, Philips). Whatever its benefits and problems it sits outside the realm of CIDOC CRM. This also extends to other recent discussions, for example, on what can have, and what is, agency. I hope this comment is a useful overview but regardless it is important that conversations are constructive and professional, and bear in mind the benefits and limitations of the CIDOC CRM. It is also high time that we recognise the enormity of the achievement! Thanks, Dominic
_______________________________________________ Crm-sig mailing list Crm-sig@ics.forth.gr http://lists.ics.forth.gr/mailman/listinfo/crm-sig