Issue 504: Formulate the philosophical underpinnings of crm and its relation to reality and the objectivity of observations.

Starting Date: 
Working Group: 

In the 47th joint meeting of the CIDOC CRM SIG and ISO/TC46/SC4/WG9; 40th FRBR - CIDOC CRM Harmonization meeting; the sig, during the discussion on issue 386, decided to open a new issue on  how to formulate the philosophical underpinnings of crm and its relation to reality and the objectivity of observations, even though the SIG has already delved in that topic –as part of the changes made in the introduction (Reality, Knowledge Bases and CIDOC CRM). But it would be nice to explore some more, even if we don’t get concrete results.

June 2020

Post by Martin Doerr (9 June 2021)

Dear All,

Here my introduction for issue 504, Philosophical Underpinnings of the CRM ,for the next meeting:

"The new official edition of the CIDOC CRM, version 7.1.1, contains a compact section with title "Reality, Knowledge Bases and CIDOC CRM". This section outlines philosophical underpinnings of the CIDOC CRM, as they emerged from addressing requirement from the community, both in terms of understanding and in terms of creating a model adequate to the scientific attitude of the user community. This paragraph does however not refer to relevant philosophical literature and theories, and does not go into further details of interest.

Therefore this issue should lead to a small working group, which will first define the philosophical question of interest, then assess priorities, collect references and finally select relevant positions supporting the way the model addresses the scientific attitude of the user community and the description of their subject matters. It should further identify counterpositions that may be popular in ontology engineering but are incompatible with the methods applied for building and using the CRM.

Since the most general topic of the CRM can be seen as the attempt to constrain and reconstruct possible pasts from evidence at our hands, and to come to a scientific agreement of reasonable and well-justified alternatives, the relationship to reality and what can be said about it, is central.

Topics of fundamental importance are, among others: reality and (numeric) identity of particulars; the concept of observer-independent reality and its relation to modelling constructs; comparability of reality with alternative modeling constructs (conceptualizations); limitations of propositional logic to describe aspects of reality (such as waves and clouds), and how facts in terms of propositional logic can relate to alternative ways of description (such as functions).

Problematic is the popular "constructivist stance" many ontologists refer to, because it ignores the "evidential constraints" Alison Wiley describes, the idea that observer independent reality limits reasonable conceptualization by empirical contradictions.

Therefore we assume that the current "scientific realism" is a better base for the CRM, however it deals with categorical knowledge produced by natural sciences and not really with the concept of historical reality of past facts in cultural, social and natural history, including geology, paleontology, biological evolution etc.

But there are also more practical issues, such as the meaning of "scientific discourse", disciplinary practices, research bias by the very model itself, and others.





Post by Francesco Beretta (Jan 6, 2022) 

Part of a discussion via the sig list that stems from the proposal to introduce properties assigning generic outcomes to activities, which can inform issue 504

Dear Martin, and (indirectly) Rob and George,

Thank you for your comments.

The point is: why care at all about building CRMsoc if everything is already present in CRMbase? And, furthermore, are we sure that we have a well grounded foundational analysis of social facts in CRMbase? Especially as we know that it's grounded in the epistemic view of classical physic (I mean not quantic)?  Isn't this a somehow different domain of discourse from social life as such? And shouldn't we take car not to model a domain without taking into account the paradigms of the disciplines studying the domain under consideration, i.e. social life?

A 'speech act' is the appellation by social philosophers of what I would call a social event from the perspective of social sciences. It can happen, of course, via a click on a website, and then it's the buyer moving the finger which is the physical,human event tht provides the setting of the acquisition speech act/social event. crm:E8 Acquisition is clearly a subclass of speech act/social event insofar as ownership is not a physical state but a social connotation added by some humans, collectively, to a physical thing. In the perspective of CRMsoc, the change in ownership, i.e. an acquisition is therefore a speech act/social event.

To come back to the specific topic, I used the example of negotiation not in a metaphysic sense but just as an example on a Saturday market. The crm:E7 Activity of type 'Negotiation about the price' of this sausage happens in time and space, but the change of ownership of the physical human made thing —once the price is set— happens just in human minds, and our there presents dogs of course still hope to receive a bit of the sausage but do not know, as non humans, what property means.

So, just for the sake of the example, and not 'deterministically' a negotiation can end in a sale or not. There are two 'events', the negociation and the sale. The type of the first event is 'negotiation' and (to use Rob's propery identifier) the 'triggered_activity_of_type' is 'acquisition' or 'non purchasing' as a E55_Type .

If we take a more physical event, an attempt of assassination of a minister, we could know who was trying the bring this activity to an end, and who was the target of the attack, but the 'oucome type' would be 'failed'.

-- Activity 'one' (of type 'one') /triggered/ (also a missing, useful property) Activity or event two (of type two).

-- [shortcut] Activity 'one' triggered activity/event of type two

The substance is in the causal sequence (as the content of the available information).

All the best


Post by Martin Doerr (Jan 7, 2022)

Part of a discussion via the sig list that stems from the proposal to introduce properties assigning generic outcomes to activities, which can inform issue 504

Please see issue 504.



Dear Francesco,

Please let us not confuse everything. The CRM is in no ways grounded in physics. It is grounded in the scholarly views of museum curators and a set of other disciplines dealing with historical facts of the past in careful collaborations over 25 years. Please study the history of the CRM, beginning from the Smithsonian in Washington. The CRM is compatible with GIS systems, physics, biology and othersciences, but does not describe their theories, but exclusively particular facts of the past. It includes social constructs already, and no no ways excludes social disciplines.

The only real bias of the CRM is purely technological and inevitable: Information integration via machines based on binary logic, as a means to link context-free identifiable individuals from human provided data, which form a small but important subset of historical data, which the a collaborative empirical result of the work of the CRM-SIG. This determines the modelling principles of the CRM.

If you introduce other functions or principles of modeling, we have first to understand their new epistemic function in the scientific discourse, and understand if and how binary logic based machines etc, may be able to process them epistemically correctly for the intended purpose. If you introduce other functions or principles of modeling, the model will necessarily be different from the CRM itself and possibly incompatible.

This *cannot* be discussed vi a e-mail exchange, it is too complex. Any such attempt would increase the already existing confusion of epistemic and  technological insights and necessities. We have to organize vitual face-to-face meetings for that purpose.

My previous messages where *solely* about the word "outcome" and its polysemy. I perfectly understand the word. I only tried to make the CRM audience aware of the ontological methodologic questions necessary to reveal the polysemy of this word and enable choosing the senses that can be modelled adequately. I also tried, without success, to make you aware of the extreme context dependency of any such word. If this has been misunderstood, we need to discuss this *face-to-face *with enough time. The necessary explanations and disambiguations definitely exceed my writing capacities.

Looking forward to a substantial face-to-face discussion and resolution of any misunderstanding,

Best wishes,



Post by Francesco Beretta (Jan 6, 2022) 

Part of a discussion via the sig list that stems from the proposal to introduce properties assigning generic outcomes to activities, which can inform issue 504

Dear Martin, 

thank you for taking of your time to discuss my statement below about the foundational perspectives of the CRM. Allow me please, in turn, to restate, with all due respect, the distinctions that we should introduce if we want to remove any misunderstanding. It would indeed be a pity to "confuse everything".

I must confess that it is almost impossible not to think that the level of factuality modelled by CRMbase is the one of the happenings in the physical world, and this from the point of view of the spatio-temporal relations of classical (Galilean or Newtonian) physics.

Of course the domain of discourse of the CRM is the one of cultural heritage and history but considered from the perspective of space-time relations of entities, i.e. of events happening in time and the physical space, as the foundational stance.

As a matter of fact all events in CRMbase are subclasses of crm:E4 Period : "This class comprises sets of coherent phenomena or cultural manifestations occurring in time and space." And, as we all know, it is precisely the crm:E4 Period class that introduces the two properties establishing a relationship of events to space (crm:P8 took place on or within → crm:E18 Physical Thing ; crm:P7 took place at → crm:E53 Place) .

Furthermore, crm:E4 Period is at the same time a subclass of E2 Temporal Entity, the class providing the time-span projection and the temporal relationships, _*and* of_ E92 Spacetime Volume, last being defined as comprising "4-dimensional point sets (volumes) in physical spacetime (in contrast to mathematical models of it, or the models of quantum mechanics) regardless their true geometric forms. They may derive their identity from being the extent of a material phenomenon or from being the interpretation of an expression defining an extent in spacetime." Space-time (physical, not mathematical) is thus central in the foundational view of the CRM.

In addition, crm:E52 Time-Span is defined as "abstract temporal extent, in the sense of Galilean physics" and crm:E53 Place as an "extent in space, in particular on the surface of the earth, in the pure sense of physics: independent from temporal phenomena and matter". Both are conceived as projections of physical phenomena into reference systems.

Finally, we can read in literature that the choice of not having phases or states, in the CRM, but just events "was inspired by considerations of modern physics, which considers Events as complex interactions with non-negligible temporal extent, leading into phases of stability with fuzzy (if any) temporal boundaries." This explains why "The CRM encourages the description of transitions rather than the states or situations in between them". Doerr Martin, Hunter Jane and Lagoze Carl, Towards a Core Ontology for Information Integration. /Journal of Digital Information/ 4 (1), 2003, 5.1 About Situations and Events (

With all this said (respectively before our eyes), it would be almost impossible, for an informed observer, not to think that classical physics provides the foundational perpective, or the fundamental insight that establishes the essential ontological categories of the CRM events and its analsys of the domain of discourse of "the scholarly views of museum curators and a set of other disciplines dealing with historical facts of the past". And that these facts are considered under the perspective of their spatio-temporal extent and characteristics.

I do not criticise this choice at all or discuss it at all as it can be considered as entirely appropriate in order to establish a ground ontology aimed at integrating factual information about cultural heritage physical objects.

Concerning conceptual objects, even the crm:E65 Creation class, producing instances of the crm:E28 Conceptual Object class, is to be considered in this view —insofar as it is a subclass of crm:E4 Period— as an event happening in space and time, as the substratum of the intellectual event happening in the mind of the artist or poet. The same applies to social events, such as the crm:E66 Formation of a group or crm:E8 Acquisition of a physical thing: it this their projection in physical reality that is considered, in application of the applied foundational view, not their social dimension and the analysis of what happens in the minds of the participants.

Thus, the point I was making, and that started our exchanges, is that if we want to model social events or phenomena *as such*, like acquisitions and ownership, joining groups and memberships, shared opinions in social networks and radicalisations, etc., i.e. as phenomena in human minds, individually or collectively, we need an extension of CRM applying a different and complementary foundational perspective.

As a matter of fact, social philosophy, social psychology and social sciences generally consider —as a standard view— that this kind of social phenomena happen in the minds and the social space as a collective space of the activities of individual human minds and as a result of individual, shared and/or collective intentionality (cf. e.g. Gallotti Mattia and Michael John (éds.), /Perspectives on Social Ontology and Social Cognition/, Dordrecht, Springer Netherlands, 2014 or Sammut Gordon et al. (éds.), /The Cambridge Handbook of Social Representations/, Cambridge, University Press, 2015).

In the CRMsoc extensions that George and myself are developing, and already proposed and will propose to discussion to the next SIG, we therefore propose to adopt the foundational perspecive of a phenomenology of social phenomena compatible with standard approaches both in social philosophy and in social sciences as an effort to carefully distinguish the physical and the social space and to add a well grounded layer of social events and states ontop of the events in the physical world that is consistent with CRMbase and its focus.

Therefore, in my perspective, "the speech act/social event is *not* identical to the crm:E8 Acquisition" (see below), the latter being the projection or setting in the physical world of the former which happens in the minds of the participants and the social space – and is not seen by our dogs. According to the distinction of foundational approaches, and the epistemological layering, the same human happening can be split into distinct physical and social events (in an appropriate extension) and this allows to capture the complexity of the happening, from different perspectives, as shown in the examples brought into the discussion by Rob and George : an archeological expedition (as social and physical endeavour) ended with 'came home empty handed' (or no death after a murder attempt): this is a specifically individual and social failure in relation to a personal/social project or plan, or an expectation by the funding agency, but not a physical event.

I consider that we should have an open discussion on ways of modelling social phenomena, and try to find a common vision in the SIG, in order to produce extensions of the CRM grounded in scientific standards that are not the same in different domains of discourse, and disciplines. And, indeed, it will be a pleasure for me to resume in the next future our exchanges, as in the past, in preparation for the next SIG.

I also kindly ask to have this post as a new issue on foundational perspectives and stances in CRM base and extentions.

All the best


Post by Martin Doerr (Jan 10, 2022)

Part of a discussion via the sig list that stems from the proposal to introduce properties assigning generic outcomes to activities, which can inform issue 504

Please see issue 504.



Post by Francesco Beretta (Jan 16, 2022) 

Part of a discussion via the sig list that stems from the proposal to introduce properties assigning generic outcomes to activities, which can inform issue 504

Dear Martin, 

thank you for the hint and reminder.

I would then kindly ask the issues-managers to add this email to issue 504 concerning : "Formulate the philosophical underpinnings of CRM and its relation to reality and the objectivity of observations. "



In the 52nd joint meeting of the CIDOC CRM SIG and ISO/TC46/SC4/WG9; 45th FRBR - CIDOC CRM Harmonization meetingMD offered a summary of the discussions that have taken place concerning the issue and made a proposal on how to proceed.

Proposal: Form a working group that works towards defining the foundational principles of the CRM and the general epistemological attitude necessary to assess them. Assuming a point of agreement has been reached, then proceed with identifying philosophical question of interest, assessing priorities, collecting references and finally with selecting relevant positions that support the way the model addresses the scientific attitude of the user community and the description of their subject matters. The group will further identify counter-positions that may be popular in ontology engineering but are incompatible with the methods applied for building and using the CRM.

FB: registered some objections regarding the objectivity of information, but is otherwise happy to proceed as MD suggested and wants to participate.

Volunteers to proceed in this line of work: FB, MD, AT, TV, EC, MA

Part of a series of excahanges through the SIG mailing list that were originally part of issue 581 (an evote), repeated here as requested by Francesco Beretta. 

Post by Francesco Beretta (25 Feb 2022; 12.29 pm)

Dear Martin, dear Franco,

I assume that the same question by Franco (Issue 581) is raised by page 25?

  • " What goes on in our minds or is 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. "
  • " priority of integrating information based on material evidence available for whatever human experience."
  • " The CIDOC CRM only commits to a unique material reality independent from the observer."

Cf. the new proposition below:

As “available documented and empirical material evidence” are regarded all types of material collected and displayed by museums and related institutions, as defined by ICOM[1], and other  collections of things providing evidence about the past, in-situ objects, sites, monuments and intangible heritage relating to fields such as social history, ethnography, archaeology, fine and applied arts, natural history, history of sciences and technology.


It seems to me that these 'fussy' questions raise in fact, once again, the relevant Issue 504 concerning the philosophical underpinnings of CRM.

The consequences of this approach are illustrated by the recently published Sealit project ontology, class: Legal Object Relationship (e.g. property of a ship by some actor): "This class comprises legal object relationships of which the timespan and the state (of these relationships) cannot be observed or documented. We can only observe these relationships through the events that initialize or terminate this state of relationship (starting event and terminating event). "

I'm not sure how many domain experts would agree with this definition because ownership of things, as a fact, is attested in written texts, or even in minds of living persons and expressed in utterances, and these are empirically observable.

The here adopted foundational stance excludes this fact (i.e. property) from being a subclass of E2 Temporal Entity.  Legal Object Relationship is declared as subclass of E1 Entity.

But on page 33 of the CRM documentation we can read: "The more specific subclasses of E2 Temporal Entity enable the documentation of events pertaining to individually related/affected material, social or mental objects that have been described using subclasses of E77 Persistent Item."

I must therefore admit that a careful reader is somewhat confused and that having an extension, such as CRMsoc, providing additional classes to deal with individual intentional and social life, and dealing with mental and social facts as empirically observable, intentional (collective) facts as we propose, could only be an advantage.

This email therefore relates to issues 504 and 580. I'd kindly ask to put it there and add there links to the relevant other issues.

All the best



Part of a series of excahanges through the SIG mailing list that were originally part of issue 581 (an evote), repeated here as requested by Francesco Beretta. 

Post by Athina Kritsotaki (28 Feb 2022; 12.25 pm) -reply to Francesco (25 Feb 2022; 12.29 pm)

Dear Francesco, dear all,

There may be a misunderstanding regarding the class Legal Object Relationship, which I explained in the presentation in the last sig meeting: We defined this class in a sense of a state of ownership of a ship, which is a kind of information that can be inferred (implicit knowledge) and not directly observed – it can be observed by the starting and terminating event of this state. It is like the soc Bond, which describes social/legal relationships that cannot be observed.

We strictly follow the modelling principle which refers that we model from actual information sources that  reveal actual practice- according to the historians of the sealit project, a ship ownership phase is described as a state with the only information documented to be about the ship owner, the shares that may have and the name of the ship, not the dates of this ownership (which is a quite complex phenomenon to observe since a person e.g may possess up to 1/48 of a ship, so you can understand how many ships shares a single person could have in the same time and there is no documented information on the timespan of this shareholding. Additionally, the ownership is used to assign a name to a ship and a ship changes its name under an ownership state. However, additional temporal information on these names under ownership states is not documented in the source – the Ownership phase can be traced by the ship registration activity (that includes timespan information) that initiates it and by the de-flagging, both events that are documented. This is material evidence, coming from the source.  If you open a Lloyd catalogue, you will find this information under ship registration without dates on the owners of the ship.

Another modeling principle that is represented in our decision to leave Legal Object Relationship as a subclass of E1 CRM Entity is that we support the progressive improvement of classification knowledge by IsA hierarchy. Since we don’t have enough knowledge and we support the open world assumption, which means that new evidence may change the classification, we prefer to model the more general (here we classified under E1) and, when we have more precise knowledge by instances on the nature of this Legal Ob.Relationship class, we can progressively specialize and refine the E1 and find the superclass under which Legal Object Relationship fits.

Sealit is a model that is based on data input, it can be refined and improved based on new knowledge, new instances.

I just wanted to explain the logic under which the model was constructed and to prove that it is one of the most representative documentations from material evidence we had, in our experience. So I am a bit confused how this use case supports raising philosophical questions regarding issue 581.


My BRs,


Part of a series of excahanges through the SIG mailing list that were originally part of issue 581 (an evote), repeated here as requested by Francesco Beretta. 

Post by Francesco Beretta (1 Mar 2022; 11.47) -reply to Athina (28 Feb 2022; 12.25 pm) 

Dear Athina,

Thank you for taking of your time and for making explicit the reasons of your modelling choices and methodology.

As University trained historians, we know that the model of the information produced by a project generally depends on the research agenda and the available sources. The model of a project is therefore not an ontology in the sense of a conceptualisation allowing for multi-project interoperability. Even the way of modelling a ship's voyage may change according to the lines of inquiry of different research projects. For this reason, a strict bottom-up modelling methodology in the field of historical research, and more broadly in the social sciences, without foundational analysis, doesn't seem to be the most appropriate way of producing an ontology for the whole portion of reality —a quite relevant portion in the cultural heritage perspective— these disciplines are concerned with.

Regarding the ownership of a ship (, which in French is in some contexts referred to under the technical term 'armement' ( — cf. "registration activity" below), the social fact of ownership is as such and in general —in the sense of ontology— observable. One can ask sailors or informed contemporaries and they will know who the owner of the ship is. There are historical sources, for example correspondence, which attest to the role of shipowners (armateurs) of such and such a person or company, even if we have lost the shipping registers which state the events of taking ownership.

In the Sealit project, a methodological choice or stance was adopted which is certainly legitimate in the project's context, but which one should avoid to generalize stating e.g. that ship ownership is not directly observable, as this would be in contradiction with observable reality. Besides the collective, attested and observable knowledge of ownership, there are, for other subdomains, written statements about it. One has to think of the land registry documents ( which often attest to the social fact of land ownership, or other rights on land, without necessarily knowing where it comes from. These rights are observable and part of reality as evidenced by the recent trials and convictions of climate activists who have occupied and organised unauthorised events at the headquarters of private companies, on the basis of infringement of private property.

So should one intend that social bonds, ownership, etc. are —in general and as such— not observable does not seem to be very prudent, because the fact of generalising a specific method of modelling, whose foundation and epistemological principles have never really been made explicit (in their foundational, philosophical aspects), risks compromising the possibility of adopting such an ontology by entire scientific communities, such as the social sciences, historical sciences, etc., whose objects are precisely related the social facts and immaterial cultural heritage.

I am therefore not at all criticizing the modelling choices of the Sealit project, which are entirely legitimate in the context of the project's model. I would simply caution against implicitly accepting foundational and philosophical modelling principles, such as those we are called to vote on —e.g. the reference to "empirical material evidence" in the context of an ontology (the CRM) that "only commits to a unique material reality independent from the observer"— regarding issues that appear to be merely about innocuous wording, and by far are not, and should actually be once explicitly formulated, discussed and accepted.

It is in this sense that I understand this question, as well as the one raised in issue 581, to fall under issues 504 and 580.

Part of a series of excahanges through the SIG mailing list that were originally part of issue 581 (an evote), repeated here as requested by Francesco Beretta. 

Post by  Martin Doerr (1 Mar 2022; 02.44 pm) reply to Francesco (3 Mar 2022; 11.47 am)

 Dear Francesco,

May I object. I maintain that ownership is not observable. All examples you provided are about memories or documents of acquisition, or about those who claim to know those (who know/have known those) who know. The events of acquisition, in whatever form, are the only one that are observable. This does not require a higher conceptual consideration in the first place. Without counterexample, I cannot follow your criticism.

All the best,


Part of a series of excahanges through the SIG mailing list that were originally part of issue 581 (an evote), repeated here as requested by Francesco Beretta. 

Post by Francesco Beretta (3 Mar 2022; 12.32) -reply to Steve  

Dear Steve,

I fear that it is reality, rather than me, that is the cause of your trouble. "a statement that an instance of Actor at a particular point in time expressed an opinion about the ownership of a vessel" is precisely the observation of a social fact, i.e. the collective belief (or disbelief) that this person is the owner of this ship. I fear that there is no other substance of ownership, as of any other social fact, that collective belief. And this is observable in human statements, be they written down or just oral.

I also fear that denying the status of observation to such an observation is neither a bottom-up approach, nor the integration of databases, nor anything else but an implicit epistemic position presented as indisputable.

Now, because what is indisputable is, by definition, not debatable, I'll stop arguing.

And take the opportunity to wish you a good day


Part of a series of excahanges through the SIG mailing list that were originally part of issue 581 (an evote), repeated here as requested by Francesco Beretta. 

Post by Martin Doerr (3 Mar 07.41 pm) -reply to Francesco Beretta (3 Mar 2022; 12.32 pm) 

Dear Francesco,

I fear no jurisdictional reality will support such undifferentiated positions, without analytical thought about who believes what, based on what, who would accept such believes with which consequences, and what kinds of thing you observe, and how this relates  to the ownership fact.

Steve, as well as I, do not present the term "observable" as an indisputable epistemic position.

We should finally understand that such e-mail exchanges cannot resolve background of years in a single message, and are therefore meaningless, if they expect absolute terms and self-consistence within a single message.

We should finally understand that the term "observable" and the substance of the observed we, Steve, Athina and me, apply, is itself a function of the particular research questions to be answered, as are ALL concepts in the CRM.

Any philosophical agreement on the general meaning of "observable" in has so far failed, as most globally valid definitions of human terms. Paul Feyerabend, in his last writings, expressed the opinion that fundamental human terms must be flexible enough in order to engulf new realities.

When Athina, Steve and me describe the ownership relation as not observable, your first question should not be questioning the prudence, but ask for what sense of "observation" we have applied. This sense, by the way, had publicly be discussed in CRM-SIG, I think in your presence and be well understood, I think, at that time.

The simple question, how someone in this society would prove his ownership of a ship, or being married, would reveal a lot of distinctions that are indisputably necessary for adequate modeling by formal ontologies based on binary logic.

Questioning the bottom-up method is even more counterproductive, because the actual sense of "social belief",  "observable fact" itself, consistent with the data and question to be answered can only be singled out on the base of bottom-up analysis, see, e.g., George Lakoff's excellent analysis of "my true mother".

Any suggestions that this my question ("how do you prove to be married") would be an expression of another simplistic assumption are quite counterproductive for the way CRM-SIG works. This question is an invitation to a methodological exercise.

My personal opinion is, if someone cannot go through at least one such an exercise in all its ramifications ("marriage witness", documents signed by, documents created based on witnessed documents, all documents lost and making claims credible, legislation changing, national archives preserving documents of witnessing legislation, acting like being married, distinguishing religious from secular authority), one can hardly claim doing generic modelling compatible with binary logic, i.e. with "formal" ontologies. Even then, "gray" fuzzy zones may remain, and need to be understood if they will affect seriously recall and precision. It is a time-consuming, exhausting and slow process, inconvenient for many, but at least the product has a reasonable long-term stability and continued extensibility.

The sense of "observability" presented by Steve, Athina and me is the one underlying the concept of being "marriage witness" or being not, as a social fact, sufficiently robust and accepted, in several relevant societies by their authorities and beyond. It is distinct from God being witness. It is distinct from observing an expression of someones opinion, and neither questions the latter, nor the way a historian would use such evidence in constructing a possible or likely past. It is no positivist threat against historical "inferences to the best explanation" from available evidence.

If other societies apply incommensurable concepts for such things, we would need again a careful analysis and understand the reasons. I remind for example David Graeber's very detailed analysis of obligation and the incompatibility of dowry with payments.

Anybody trying other intellectual methods is kindly invited to follow that and then show if it helps answering the respective research questions, to explain the data, and to produce the best automated inferences.



Part of a series of excahanges through the SIG mailing list that were originally part of issue 581 (an evote), repeated here as requested by Francesco Beretta. 

Post by  Martin Doerr (1 Mar 2022; 04.31 pm) reply to Francesco (3 Mar 2022; 12.32 pm)

Please let me further add, that custody in the sense of the CRM is observable. Any legal transition from custody to ownership will again be based on observable proof of custody.

Please also respect that the Sealit model has passed practice test with a huge amount of data under the scrutiny of enough international experts in maritime history from different countries. Following the way we use to argument in CRM-SIG, claims about lack of genericity, or inadequacy for other applications, of any local model as the Sealit model, must always be based on instance data and explicit formulation of the research question not answered.

If the research questions require incommensurable modeling, such as the classical difference between molecular theory and thermodynamics, no theoretical top down framework can unite them. We have successfully tried since 1996 to push the limits of commensurable modeling widely across context and disciplines, by a strict intellectual discipline regarding the kinds of reaserch questions that can be answered correctly. We never excluded the justified existence of incommensurable models.

General claims of incompatibility of domain models circulate since decades. All the enterprise of formal ontologies was to overcome it. No top-down theoretical framework since then has produced a viable result since then. The epistemological insight, which sort of research questions allow for an increasing set of cross-disciplinary compatible models, is the most demanding scientific endeavour of CRM-SIG.

Theoretical understanding is vital also for bottom up modelling. Otherwise, no reasonable candidate constructs can be found. But, no construct, regardless of how abstract, should be accepted without practice test. CRM-SIG is committed to producing international recommendation. If Sealit is in the practical scope of, e.g., CRMsoc, and its research questions, at some point in time, Sealit must be be integrated by practice test.

All the best,


Part of a series of excahanges through the SIG mailing list that were originally part of issue 581 (an evote), repeated here as requested by Francesco Beretta. 

Post by  George Bruseker (1 Mar 2022; 04.41 pm) -reply to Martin Doerr (1 Mar 2022; 04.31 pm) 

Dear all,

Social symbolic events such as acquisitions (not done by force) are also strictly not observable since you can only know that they occur if you share the same social symbolic set and 'conclude' or 'infer' that something has taken place. There is no atomic level at which we see these things and can then say 'and now it is done'! Which atom, at what moment? Of course there are various pieces of evidence you can go looking for and say these are the things you must observe, but it's an obtuse way of looking at things because if you are at the wedding and you are a literate member of the cultural group then you know (barring an evil demon) that when the bride has been kissed (and some books signed) that the event has occured. You 'observed' it.

It is reasonable and natural for how to structure information and how to ask questions to posit an observation acquisition event rather than saying that what is observable is the book, the handshake etc.

This is the same with social institutions. No document need be consulted for an alien anthropologist to land amongst CRM SIG discussion and determine who the leader is. Having read a few background documents about general human culture and observing a set of behaviours amongst a group of people the anthropologist 'observes' M Doerr to be the leader. To say that this is not observable is extremely hard to support (except again if we argue only atomic configurations can be observed). What was observed is not necessarily initiating and ending events (also symbolic, also only knowable beyond physical material evidence), but a number of indicators within a social symbolic system which indicated this to be the case.

It is thus equally natural to say that the social fact is observed although in fact many minute individual observations were made etc. It would be obtuse to ask for these to be listed instead of the fact in the same way it would be for the event because this is not the form of evidence that is typically required in the domain on inquiry.

Francesco points out for the nth time, and I'm not sure why this cannot be heard or acknowledged, that historians usually do not have the kind of evidence you ask for of physical events in space and time that start social states. The historian is not at fault, the historical record is imperfect. It is in this case not for the historian to change his practice but for the ontologist to provide a structure which relates to the kind of reality that the expert tries to describe.

As in observation in the sense of physics, the observer can be wrong.



Meetings discussed: