The paper presents work in progress towards an applied ontology for Digital History. The challenge of ontology-based modeling for historical understanding is to make explicit the conceptualizations of historians. As philosophy of history shows, historians build conceptual models of historical events by the method of 'colligation'. Grasping different points of view on one and the same historical event is an essential requirement of historical research. Hence, I focus on the representation of multiple perspectives on historical events. Although being already well-established in the biomedical and legal domain, philosophically informed ontology design principles and patterns are rarely applied in the Digital Humanities. Thus, the paper explores the role of philosophy and the relevance of the widely neglected use of top-level foundational ontologies, ontology design patterns, and Semantic Web technologies for augmenting historical understanding via knowledge modeling.
The representation of historical reality as seen from the divergent perspectives of
different historians, as perceived by historical actors or as reported in historical sources has
requirements beyond the straightforward ‘reality representation’ [
Despite the important role of philosophical ontology for a “coherent conception of historical knowledge” [30], the applied ontology approach is widely neglected in the domain of Digital History and in the Digital Humanities in general. In line with Garbacz and Trypuz [19], who state that, “[i]f neither the standard nor the simple patterns meets your needs, we think you will have to consider a solution that involves some philosophy”, I propose an approach towards ontology-based modeling of historical knowledge informed by philosophy of history. The goal is to provide an applied ontology approach to support historians in modeling their expert knowledge about historical events. I suggest a modular architecture reusing and extending Ontology Design Patterns (ODPs) [25] for knowledge modeling tasks in the domain of (Digital) History.
Work—theoretically very similar to the philosophically informed knowledge modeling
approach presented in this paper—was already introduced by Shaw [
The Spatial History Ontology (SHO) developed by Grossner [21] applies the
toplevel ontology DOLCE [18] and its Descriptions and Situations (DnS) ODP [17] and thus
enables the event-based modeling of multiple interpretations of historical processes
including spatial information. An alternative to the DnS pattern is the Multiple
Interpretation Data Model (MIDM) [
In summary, regarding the ontologies which explicitly state the feature of representing multiple perspectives, views, or interpretations of historical events, SEM is suited to represent multiple perspectives of different actors involved in historical events (second use case addressed in the introduction) and views according to different historical sources (third use case) via its sem:accordingTo property, but not to grasp conceptualizations of whole historical processes as conceived by different historians (first use case). In contrast, MIDM allows the modeling of different conceptualizations of historical processes 1See LODE ontology: http://linkedevents.org/ontology/ 2The life cycle of historical information [7] is a good tool to locate the use of Semantic Web technologies in Historical Information Science [33]. Ontologies are typically applied in the enrichment and editing phases [cf. 33, p. 10], but not in the analysis phase of the life cycle. Though, using Eide’s distinction between ‘modeling for production’ and ‘modeling for understanding’ [13], I focus on the task of knowledge modeling as modeling for (historical) understanding, not as modeling for production (e. g. the production of digital editions enriched with knowledge graphs to support information retrieval and exploration of the content).
3SEM is not aligned with a foundational ontology, has no properties to model relationships between events and has no thematic roles. (M5 Sequence class), but neglects the modeling of roles. Despite its advanced representation of thematic and social roles, HERO does not yet provide a DOLCE extension to support perspectival role modeling. Regarding the other DOLCE-based ontologies, both SHO and F use the DnS pattern—(F via DUL4)—and therefore are well prepared for modeling multiple conceptualizations of historical events. However, SHO’s emphasis is on the geospatial domain and F was designed with distributed event-based system applications in mind.
Historians perform conceptualization in order to understand historical events. Philosophers of history analyzed this activity as colligation. Walsh [53, p. 59] defines colligation as “the procedure of explaining an event by tracing its intrinsic relations to other events and locating it in its historical context”.
According to Abell [1, p. 39 f.] there are two meanings of colligation: tracing causal connections between events and classifying interconnected events. Though colligation can also be seen as a two-step process or procedure consisting of these two tasks (see UML Use Case diagram in fig. 1).
Ontology-Based Historical Information System
Ontology-Based Historical Information System Historian
Trace relations between events Classify interrelated events
Trace chronological relations Trace mereological relations Trace causal relations
The Descriptions and Situations ODP (DnS) [17] is applied to support the process of colligation (see fig. 2) by means of ontology-based modeling as sense-making tool [see also 14].5 A Description represents the colligatory concepts and relations grasped in a 4http://ontologydesignpatterns.org/wiki/Ontology:DOLCE+DnS_Ultralite 5DnS is not limited to DOLCE and can be adjusted to other top-level ontologies with the categorial distinction between endurants and perdurants. CRM has the categories E77 Persistent Item for endurants “synoptic judgment” [34]. In bio-medical ontologies Descriptions are used to represent medical diagnoses [e. g. 16]. There is indeed an interesting analogy between the synoptic judgments of a historian and medical diagnoses of a physician: a combination of broad medical knowledge, relevant evidence drawn from various tests, a knowledge of various theoretical possibilities for explanation, and skill in seeing which interpretation of the evidence works best in a particular case—the difference being, of course, that the physician deals primarily with law-bound physiological processes, the historian primarily with human conduct and purposive action. [46, p. 69]
A Situation represents the explanatory relevant historical context, i. e. historical events and their interrelations. The intent of the Situation pattern is to “represent contexts or situations, and the things that are contextualized”6. A Description “provides an interpretation to a set of observed entities”7.
Hence, a Situation is a historical context of related events and entities as interpreted in a synoptic judgement (represented by a Description). Note that an instance of a historical process in our ontology is not a process (as perdurant) [see also 21, p. 169], but a concept conceptualized by a historian (e. g. the colligatory concept ‘Russian Revolution’). In this way, it is possible to represent multiple perspectives on one and the same historical phenomenon by modeling Descriptions which define colligatory concepts as expressed in different historical narratives (see lmm:expresses in fig. 2).
To summarize the proposed approach towards an applied ontology for knowledge modeling in Digital History, the following paragraphs provide an overview of the required modules or components which cannot be discussed in-depth within this paper. Event-based Modeling enables the representation of temporal, mereological, and causal or constitutive relations between events.
Role-based Modeling enhances event-based modeling with thematic and social roles played by the agents involved in events.
Levels of Reality to represent entities on different ontological levels (e. g. political and economical in the social sphere) [see 38] can be modeled with DnS in line with the stratification principle and reification principle [cf. 15, with examples from the legal domain]. (See example for epistemological layering in the biomedical domain by Gangemi, Catenacci, and Battaglia [16, diagram in fig. 6].) Semiotics has to be added to the ontology to further clarify the representation of historical reality. Indeed there is some confusion in the distinction between representation and the represented with the distinction between representation and reality in the DnS pattern used without explicit modeling of semiotic relations [cf. 26, p. 27]. Different interpreta(agents and objects) and E2 Temporal Entity for perdurants (events) and therefore can be extended with an ontology module for the DnS pattern.
6http://ontologydesignpatterns.org/wiki/Submissions:Situation 7http://ontologydesignpatterns.org/wiki/Submissions:Description dul:Parameter dul:Role
hpr:HistoricalProcessCourse dul:parametrizes dul:isRoleOf
hpr:relates
Relational Context
dul:de nes
dul:satis es
dul:Description
dul:Situation
dul:Region
dul:Agent
dul:Event
hpr:HistoricalProcessSituation
dul:hasSetting
dul:isAgentIncludedIn
dul:isEventIncludedIn
tions of one and the same historical event or historical source can be explicitly modeled
with the Semiotic Ontology Design Pattern [5] or the Linguistic Meta-Model (LMM) in
addition to DnS. LMM’s basic classes Expression, Meaning, and Reference are the
building blocks for Peircean semiotic triangles [see diagram in fig. 2 in 37].
Frames are considered as knowledge patterns and can be used to validate ODPs [
I have shown that the classification of historical events by colligatory concepts and the multi-perspectival representation of historical reality as seen from the ‘Sehepunckte’9 of different historians, conceived in their synoptic judgements and presented in historical narratives can be well supported by applying the DnS and LMM ODPs.
Special ODPs have to be created following a modular approach in order to model causal narratives, conflict trajectories, biographies, historical travelogues, etc. for specific case scenarios and use cases (e. g. modeling multiple causal narratives for compara8Compare for example the Travel frame in FrameNet with the Linked Places project’s conceptual model of historical geographic movement: http://commons.pelagios.org/2016/10/ linking-linked-places-project-update/
9See Chladenius’s [10, p. 100] concept of viewpoint (Sehepunckt) in his theory of multiperspectivity: http: //www.deutschestextarchiv.de/book/view/chladni_geschichtswissenschaft_1752/?p=136 tive historical analysis or modeling divergent conflict histories as seen from the perspectives of conflict parties or as reported in different historical sources10). So, the next steps are the further development and the evaluation of the approach in selected case studies.11 Furthermore, phenomenological ontology as per Husserl’s idea of formal ontology [27] will be applied to analyze and conceptualize the constitution of objects of historical research in different realms and on different layers of reality.12
Digital historians should be able to formalize their expert knowledge—at least if the
high demand of philosophy of history is taken seriously by considering that “the more
schematic the conceptualization in a discipline, the more its practitioners are likely to
engage with models rather than concepts” [31, p. 25]. Nevertheless, there remains the
challenge of the computational requirements of “total explicitness and absolute
consistency” [31, p. 5] [see also 36, p. 1099] due to the problem of the ‘semantic gap’ between
narrative descriptions and ontology-based models of historical processes [see also 32].
As Saab and Fonseca [
Acknowledgements This work was partially supported by a visiting research fellowship at the Trier Center for Digital Humanities (TCDH) in November and December 2018.
10The Historical Context Ontology (HiCO) [11], for example, provides advanced modeling provencance and interpretations of content and context information.
11Campbell [8, 9] presents a chronology of 32 historical events identified and extracted from 10 different
narratives about the Bosnian War. His review and analysis of the 10 monographs opens up a straightforward
case study in modeling: I consider this chronology as (chronicle-like) ‘background narrative’ in the sense of
Bergmann [see 6, p. 58]. Ontology-based modeling with the DnS pattern will show how these events are
conceptualized and interrelated in the controversial perspectives as presented in the different narratives (see fig. 2).
Furthermore, it provides an interesting example for a critical examination of perspectival explanation (see the
critique of Campbell’s Nietzschean perspectivism in his MetaBosnia by Wight [
12GFO (General Formal Ontology) would already allow to model strata (material, mental, social) and different layers within these by default [24].
13See [
Digital Libraries and Multimedia Archives. Springer, 2017, pp. 112–121. [5] Neil Benn et al. “Ontological Foundations for Scholarly Debate Mapping Technology”. In: Computational Models of Argument: Proceedings of COMMA 2008.
IOS Press, 2008, pp. 61–72. [6] Klaus Bergmann. Multiperspektivita¨t: Geschichte selber denken. 3rd ed. Methoden historischen Lernens. Schwalbach/Ts.: Wochenschau Verlag, 2016. [7] Onno Boonstra, Leen Breure, and Peter Doorn. “Past, present and future of historical information science”. In: Historical Social Research 29.2 (2004), pp. 4–132. [8] David Campbell. “MetaBosnia: Narratives of the Bosnian War”. In: Review of International Studies 24.2 (1998), pp. 261–281. [9] David Campbell. “Ontopology: Representing the Violence in Bosnia”. In: National Deconstruction: Violence, Identity, and Justice in Bosnia. Minneapolis: University of Minnesota Press, 1998. Chap. 3, pp. 33–81. [10] Johann Martin Chladni. Allgemeine Geschichtswissenschaft – worinnen der Grund zu einer neuen Einsicht in allen Arten der Gelahrheit geleget wird. vollsta¨ndige digitalisierte Ausgabe. Leipzig: Deutsches Textarchiv (Kernkorpus), 1752. [11] Marilena Daquino and Francesca Tomasi. “Historical Context Ontology (HiCO): A Conceptual Model for Describing Context Information of Cultural Heritage Objects”. In: MTSR 2015: Metadata and Semantics Research. Vol. 544. Communications in Computer and Information Science. 2015, pp. 424–436. [12] Martin Doerr. “The CIDOC CRM: An Ontological Approach to Semantic Interoperability of Metadata”. In: AI Magazine 24.3 (2003), pp. 75–92. [13] Øyvind Eide. “Ontologies, Data Modeling, and TEI”. In: Journal of the Text Encoding Initiative 8 (2014). [14] Thomas C. Eskridge and Robert Hoffman. “Ontology Creation as a Sensemaking
Activity”. In: IEEE Intelligent Systems 27 (5 Oct. 2012). [15] Aldo Gangemi. “Design patterns for legal ontology constructions”. In: Proceedings of the 2nd Workshop on Legal Ontologies and Artificial Intelligence Techniques. Ed. by Pompeu Casanovas et al. 2007, pp. 65–85. [16] Aldo Gangemi, Carola Catenacci, and Massimo Battaglia. “Inflammation Ontology Design Pattern: An Exercise in Building a Core Biomedical Ontology with Descriptions and Situations”. In: Ontologies in Medicine. Ed. by Domenico M.
Pisanelli. Vol. 102. Technology and Informatics. IOS Press, 2004, pp. 64–80. [17] Aldo Gangemi and Peter Mika. “Understanding the Semantic Web through Descriptions and Situations”. In: On The Move to Meaningful Internet Systems 2003: CoopIS, DOA, and ODBASE. Ed. by Robert Meersman, Zahir Tari, and Douglas C. Schmidt. Springer Berlin Heidelberg, 2003, pp. 689–706. [18] Aldo Gangemi et al. “Sweetening Ontologies with DOLCE”. In: Knowledge Engineering and Knowledge Management: Ontologies and the Semantic Web. Ed. by Asuncio´n Go´mez-Pe´rez and V. Richard Benjamins. Berlin, Heidelberg: Springer Berlin Heidelberg, 2002, pp. 166–181. [19] Paweł Garbacz and Robert Trypuz. “Representation of tensed relations in OWL: A survey of design patterns”. In: Research Conference on Metadata and Semantics Research. Springer, 2017, pp. 62–73. [20] Anna Goy, Diego Magro, and Marco Rovera. “On the role of thematic roles in a historical event ontology”. In: Applied Ontology 13.1 (2018), pp. 19–39. [21] Karl Grossner. “Representing historical knowledge in geographic information systems”. PhD thesis. University of California, Santa Barbara, 2010. [22] Thomas R. Gruber. “A translation approach to portable ontology specifications”.
In: Knowledge Acquisition 5.2 (1993), pp. 199–220. [23] Willem R. Hage et al. “Design and use of the Simple Event Model (SEM)”. In: Web Semantics: Science, Services and Agents on the World Wide Web 9.2 (2011), pp. 128–136. [24] Heinrich Herre. “Formal Ontology and the Foundation of Knowledge Organization”. In: Knowledge Organization 40.5 (2013), pp. 332–339. [25] Pascal Hitzler et al., eds. Ontology Engineering with Ontology Design Patterns: Foundations and Application. Vol. 25. Studies on the Semantic Web. IOS Press, 2016. [26] Rinke Hoekstra et al. “LKIF Core: Principled Ontology Development for the Legal
Domain”. In: Law, Ontologies and the Semantic Web. IOS Press, 2009, pp. 21–52. [27] Edmund Husserl. First Book: General Introduction to a Pure Phenomenology.
Ideas Pertaining to a Pure Phenomenology and to a Phenomenological Philosophy.
The Hague: Martinus Nijhoff, 1983. [28] Anthony K. Jensen. “Positivism and perspectivism”. In: Nietzsche’s Philosophy of
History. Cambridge: Cambridge University Press, 2013. Chap. 5, pp. 119–154. [29] Ewa Kowalczuk and Agnieszka Ławrynowicz. “The Reporting Event Ontology Design Pattern and Its Extension to Report News Events”. In: Advances in Ontology Design and Patterns. Vol. 32. Studies on the Semantic Web. 2017, pp. 105– 117. [30] Daniel Little. “Historical Concepts and Social Ontology”. In: New Contributions to the Philosophy of History. Vol. 6. Methodos. Springer, 2010, pp. 41–72. [31] Willard McCarty. Humanities Computing. Basingstoke: Palgrave Macmillan, 2005. [32] Alexander Mehler and Andy Lu¨cking. In: Computational Humanities – Bridging the Gap Between Computer Science and Digital Humanities. Ed. by Chris Biemann et al. Vol. 14301. Dagstuhl Seminar. Schloss Dagstuhl – Leibniz-Zentrum fu¨r Informatik, 2014, pp. 91–92. [33] Albert Meron˜o-Pen˜uela et al. “Semantic Technologies for Historical Research: A
Survey”. In: Semantic Web 6 (2014), pp. 539–564. [34] Louis O. Mink. Historical Understanding. Ithaca: Cornell University Press, 1987. [35] Louis O. Mink. “Narrative Form as a Cognitive Instrument”. In: The Writing of History: Literary Form and Historical Understanding. Ed. by Robert H. Canary and Henry Kozicki. Madison: University of Wisconsin Press, 1978, pp. 129–149. [36] Ruth Mostern and Ian Johnson. “From named place to naming event: creating gazetteers for history”. In: International Journal of Geographical Information Science 22.10 (2008), pp. 1091–1108. [37] Davide Picca, Alfio Massimiliano Gliozzo, and Aldo Gangemi. “LMM: an OWLDL MetaModel to Represent Heterogeneous, Multilingual Lexical Knowledge”. In: Proceedings of the International Conference on Language Resources and Evaluation, LREC 2008. European Language Resources Association (ELRA), 2008. [38] Roberto Poli. “The Basic Problem of the Theory of Levels of Reality”. In: Axiomathes 12.3-4 (2001), pp. 261–283.