Mining and Modeling Spaces and Places for Literary History as Linked Open Data

paper, specified "short paper"
Authorship
  1. 1. Julia Röttgermann

    Universität Trier

  2. 2. Maria Hinzmann

    Universität Trier

  3. 3. Katharina Dietz

    Universität Trier

  4. 4. Henning Gebhard

    Universität Trier

  5. 5. Anne Klee

    Universität Trier

  6. 6. Johanna Konstanciak

    Universität Trier

  7. 7. Christof Schöch

    Universität Trier

  8. 8. Moritz Steffes

    Universität Trier

Work text
This plain text was ingested for the purpose of full-text search, not to preserve original formatting or readability. For the most complete copy, refer to the original conference program.


Introduction
In literary history and historiography, places and spaces play an important role – not least in the context of the ‘spatial turn’ (Lafon, 1997; Piatti et al., 2009; Dennerlein, 2009; Weber, 2014). In literary works, narrative locations are particularly relevant, but places of publication as well as further spatial dimensions can also be taken into account (Curran 2018; Burrows et al. 2016). Our contribution presents how we obtained spatial statements from three different information sources and combined them in a knowledge network based on the Linked Open Data (LOD) paradigm (Berners-Lee, 2006; Hooland und Verbough, 2014; Hitzler, 2021).

Project context
The aim of the project Mining and Modeling Text is to establish an information network for the humanities built from various sources.
See https://mimotext.uni-trier.de/en/.
This aim is closely linked to the finding that, considering the steadily growing digital cultural heritage, the acquisition of knowledge from large amounts of text and data can no longer be handled by individuals. In representing knowledge as LOD, we see untapped potential that we are exploring in the current project phase on the French Enlightenment novel (Delon/Malandain, 1996; Mylne, 1981).

Creating spatial statements
Information on spatial statements relevant to our domain is extracted from three different types of sources: 1. bibliographic metadata, 2. primary sources, 3. scholarly publications. After focusing on thematic statements in an earlier phase of our project (Schöch et. al., 2022; Röttgermann et al., 2022), we are currently addressing spatial statements.

Mining
Bibliographic metadata: For our domain, the Bibliographie du genre romanesque français 1751-1800 (BGRF, Mylne et al., 1977) is central, as it defines the population of about 2000 French Enlightenment novels. The BGRF has been extensively analysed and modeled (Lüschow 2020) and contains rich metadata (including places of publication, narrative locations, narrative form, characters, themes, style).
Primary sources: The Collection of Eighteenth-Century French Novels (Röttgermann, 2021) is analysed via SpaCy’s (Honnibal und Montani, 2017) named entity recognition and reconciliation pipeline supported by OpenRefine (Huynh, [2012] 2010). Our pipeline requires human intervention (Hinzmann et al., 2022) concerning the challenges of ambiguity, fictionality and historicity (Heuser et al., 2016; Jockers, 2016; Nielsen, 2016).

Figure 1: Reconciliation of narrative locations with OpenRefine

Modeling
Our approach relies on importing both the text strings as found in the information source (green) and the abstract spatial items (blue) into our Wikibase instance. Combined with the fact that we reference all statements uniquely via the "stated in" property (orange), this ensures a high degree of verifiability of our data (see fig. 2).

Figure 2: ‘Narrative locations’ of Diderots La religieuse from NER and BGRF data
Our spatial vocabulary was built up incrementally and provides items (=spatial concepts) for 7 properties, of which 5 are currently mapped with Wikidata (see fig. 3).

See for the vocabulary: https://github.com/MiMoText/vocabularies/blob/main/spatial_vocabulary.tsv.

Figure 3: Ontology of ‘spatial statements’ within the domain of literary history/historiography

Infrastructure
For the provision of data, we follow Open Science principles, such as the publication of FAIR data in open access as well as the use of open source software – in particular Wikibase (see fig.4).

Generally, see Suber 2012 and Wilkinson et al. 2016; related to the project, see Röttgermann and Schöch 2020 and Schöch, 2021. We expect the Wikibase instance to become publicly available in mid-2022: https://www.mimotext.uni-trier.de/en
We created a custom bot using the Python library Pywikibot to import and update the RDF triples into our Wikibase instance from TSV files.

See https://github.com/MiMoText/Wikibase-Bot.

Fig. 4: Local Wikibase instance

Spatial Querying
Having all the spatial triples stored in our Wikibase allows us to query and visualize them using the DockerWikibaseQueryService interface. We can gain an overview of the entire set of metadata (see fig. 5, query 1), see places of publication appearing and disappearing over time (see fig. 6, query 2) or explore narrative locations linked to specific thematic concepts such as ‘miracle’ (see fig. 7, query 3). Via 'federated queries' (see fig. 8, query 4), information from other knowledge bases (here Wikidata) can be used.

Figure 5: Overview of the most frequent places of publication

Figure 6: Places of publication over time

Figure 7: Narrative locations linked to the thematic concept ‘miracle’ (excerpt)

Figure 8: Cluster of dominant publication places based on a federated query (Wikidata)

Conclusion
We showcase key aspects of spatial information extraction and modeling in our project. Our presentation will show how we link spatial statements from three sources into a multilingual knowledge network. Triples on publication dates, themes, locations and authors can be combined and be differentiated by their source, something which allows new perspectives for literary history, book history and other domains. Future work concerns extracting and adding statements from scholarly publications to the Wikibase instance.

Appendix

Query 1: Items (novels) and their place of publication (fig. 5)

Query 2: Places of publication over time (fig. 6)

Query 3: Narrative location of novels with theme “miracle” (fig. 7)

Query 4: Places of publication with geocoordinate location via federated query (fig. 8)

Bibliography

Berners-Lee, T. (2006).
Linked Data – Design Issues. https://www.w3.org/DesignIssues/LinkedData.html.

Burrows, S. et al. (2016). Mapping Print, Connecting Cultures.
Library & Information History, 32(4), pp. 259–71. 10.1080/17583489.2016.1220781.

Curran, M.
(2018).
The French Book Trade in Enlightenment Europe I: Selling Enlightenment. London: Bloomsbury Publishing.

Delon, M. and Malandain, P. (1996).
Littérature française du XVIIIe siècle. Paris: Presses universitaires de France. https://gallica.bnf.fr/ark:/12148/bpt6k48060529.

Dennerlein, K. (2009).
Narratologie Des Raumes. Berlin, New York: De Gruyter.

Heuser, R., Algee-Hewitt, M. and Lockhart, A. (2016). Mapping the Emotions of London in Fiction, 1700–1900: A Crowdsourcing Experiment. In
Literary Mapping in the Digital Age. London: Routledge.

Hitzler, P. (2021). A Review of the Semantic Web Field.
Commun. ACM. 10.1145/3397512.

Honnibal, M. and Montani, I. (2017).
SpaCy 2: Natural Language Understanding with Bloom Embeddings, Convolutional Neural Networks and Incremental Parsing.

Hooland, S. van. and Verborgh, R. (2014).
Linked Data for Libraries, Archives and Museums: How to Clean, Link and Publish Your Metadata. Facet Publishing.

Huynh, D. (2010).
OpenRefine. OpenRefine. https://github.com/OpenRefine/OpenRefine.

Jockers, M. L. (2016). The Ancient World in Nineteenth-Century Fiction; or, Correlating Theme, Geography, and Sentiment in the Nineteenth Century Literary Imagination
. Digital Humanities Quarterly, 010(2).

Lafon, H. (1997).
Espaces Romanesques Du XVIIIe Siècle, 1670-1820: De Madame de Villedieu à Nodier. 1. éd. Paris: Presses Universitaires de France.

Martin, A., Mylne, V. and Frautschi, R. L. (1977).
Bibliographie du genre romanesque français, 1751-1800. London: Mansell.

Mylne, V. (1981).
The Eighteenth-Century French Novel: Techniques of Illusion. 2nd Edition. Cambridge, New York: Cambridge University Press.

Nielsen, F. A. (2016). Literature, Geolocation and Wikidata. In
Wiki@ICWSM. Proceedings of the International AAAI Conference on Web and Social Media. Cologne, pp. 61–4. https://ojs.aaai.org/index.php/ICWSM/article/view/14833.

Röttgermann, J. (ed.) (2021). Collection de romans français du dix-huitième siècle (1750-1800) / Eighteenth-Century French Novels (1750-1800) [dataset].
Release v0.2.0. 10.5281/zenodo.5040855.

Schöch, C. et al. (2022). Smart Modelling for Digital Literary History.
IJHAC: International Journal of Humanities and Arts Computing [Special Issue on Linked Open Data], 16(1), pp. 78–93. https://doi.org/10.3366/ijhac.2022.0278.

Suber, P. (2012).
Open Access. Cambridge, Mass: The MIT Press.

Weber, A.-K. (2014).
Mapping Literature: Spatial Data Modelling and Automated Cartographic Visualisation of Fictional Spaces. Zurich: ETH Zurich. 10.3929/ETHZ-A-010106067.

Wilkinson, M. D. et al. (2016). The FAIR Guiding Principles for Scientific Data Management and Stewardship.
Scientific Data, 3(1), p. 160018. 10.1038/sdata.2016.18.

If this content appears in violation of your intellectual property rights, or you see errors or omissions, please reach out to Scott B. Weingart to discuss removing or amending the materials.

Conference Info

In review

ADHO - 2022
"Responding to Asian Diversity"

Tokyo, Japan

July 25, 2022 - July 29, 2022

361 works by 945 authors indexed

Held in Tokyo and remote (hybrid) on account of COVID-19

Conference website: https://dh2022.adho.org/

Contributors: Scott B. Weingart, James Cummings

Series: ADHO (16)

Organizers: ADHO