FPO: Factoid Prosopography Ontology

The Factoid Prosopography Ontology (FPO) is available from GITHub. It is an OWL/RDFS based semantic web ontology that defines components for the factoid approach to prosopography as used by the Department of Digital Humanities at King’s College London (DDH/KCL) in a formal manner.  It can be used to provide a basis for a data structure for any particular prosopographical project that wishes to make use of the factoid approach. The FPO ontology is most definitely work in progress rather than a finished product.  It is presented here to promote thinking about a more formal foundation for the factoid approach.

Why create a formal ontology?

Why might FPO be useful? It formally defines components of the structure behind factoid prosopography that has worked for a number of prosopographical projects.  It thus might usefully provide a basis for a data structure for any particular prosopographical project that wishes to make use of the factoid approach, and will draw attention to structures that might not initially come to mind to a designer.

Furthermore, if two prosopographical projects are created based on a shared ontology, they are more likely to be compatible if their data need to be brought together.  It may not be an entirely trivial matter to, say, bring DDH/KCL’s People of Medieval Scotland together with People of Northern England, but it is likely to be significantly easier than it would be if the two datasets did not share a significant amount of structure between them.

FPO and existing Factoid Prosopographies

FPO and this website about it has been worked on by John Bradley at DDH/KCL because the factoid model came first from KCL, and has been used successfully as the basis for a number of different prosopographical projects there. However, the relationship between FPO and these projects is not entirely straightforward. First, all these prosopographical projects used relational technology as the basis for their data storage, and were all created before FPO was defined in 2017. Furthermore, each of the DDH./KCL project’s data structure is somewhat different from the other “sister” factoid prosopographical structures, reflecting not only the different interests of our historian partners, but also the growing understanding over time within DDH of what seemed to us to be the best way to think about the factoid model. Thus, while a “core” of FPO is drawn from the practice of CCH/DDH in its factoid prosopographies, none of them are fully represented by FPO. Nonetheless, to some degree FPO as presented here still does represent structured data as it is actually modelled in the various prosopographical projects that were developed by CCH at KCL during the 1990s and 2000s and into the 2010s, and FPO can be thought of as how we now think of the factoid model in 2017, after more than 20 years of working on different factoid-based projects.

How does the FPO relate to other ontologies?

References to TEI dates, CIDOC-CRM and FRBRoo appear in FPO, but their place in FPO’s model in fact does not represent actual practice in CCH’s factoid prosopographies.  They are here to provide suggestions, as it were, about how the factoid model might connect with CIDOC and FRBR based data, and with TEI’s representation of historical dates. Connections between FPO and other established ontologies, such as Friend of a Friend (FOAF) for example, might come to mind to the reader, and FPO’s semantic web ontology frameworks make it possible to simply extend FPO to include these kind of connections.

About CIDOC-CRM: Some preliminary discussions with folk involved in CIDOC-CRM a few years ago centred on the idea that much more of what is presented now in FPO (indeed, perhaps all of it!) were already represented in CIDOC-CRM. I’d be interested in a more thorough investigation of this idea, although I haven’t yet had the opportunity to develop an in-depth understanding of CIDOC-CRM to carry out this analysis myself.


Tailoring the FPO for an individual project

The FPO is most useful to the extent that its model is used as the basis for other factoid-oriented prosopographies.  How constraining need this be?

The fact that FPO does not represent any particular DDH/KCL prosopographical project shows that the FP ontology need not be the definitive model for how factoid prosopography must be done. Indeed, FPO focuses only an a few primary concerns about the factoid model, and any prosopographical project must expect to need to at least extend FPO to meet their needs and interests.  FPO provides, at best, a core set of structures upon which an actual prosopography could build.

Semantic Web ontology technologies are, in fact, designed in ways that are premised on the idea that any real dataset is likely to need to extend a SW  model to meet specific requirements of the dataset.  Thus, it is entirely standard practice to extend components in FPO by defining specialised subclasses or subproperties.  This mechanism can be used to, for example, link FPO’s persons to the well-established Friend of a Friend (FOAF) ontology. Standard ontology extension practices could be used to define a new kind of factoid, etc etc. Classes in a project’s prosopographic structure will almost certainly need to add more attributes to the relatively few defined for FPO’s classes.  Naming, for example:  FPO defines a head-name for historical persons, but some project might need a more complex name structure to reflect semantic structure in the naming conventions of the society they are working with. Furthermore, certain parts of the FPO might be completely ignored for a particular dataset. From the perspective of the Semantic Web at least, changes of these kinds still result in data that is largely compatible with any other dataset that applies the FPO model, and makes it relatively easy to bring two such datasets together into a single repository when it might be useful to do so.

Changes that alter FPO relationships or that significantly alter the meaning of FPO terminology result in datasets that do not fit together so easily with other FPO-compatible data.  Of course, if any such fundamental variance from FPO is required by the nature of the material being represented, than this incompatibility will have to be thought of simply as a key characteristic of this non-conforming dataset.

Where is the FPO Formal Ontology stored?

The FPO formal ontology is a light-weight OWL and RDFS based ontology.  Its OWL file is on GitHub at https://github.com/johnBradley501/FPO. In addition, some explanatory material to complement the OWL file can be found on the current site.  Look here.