Center for Uncertainty Studies Blog

Digital Academy 2023: Reflections on uncertainty in a mixed methods study on Heimat, Affect, and Whiteness

Center for Uncertainty Studies Blog — Mon, 1 Jul 2024 08:08:14 +0200

From September 25 to 28, 2023, the Digital History Working Group at Bielefeld University welcomed participants to the Digital Academy, themed "From Uncertainty to Action: Advancing Research with Digital Data." This event delved into the complexities of data-based research, exploring strategies to navigate uncertainties within the Digital Humanities. In a series of blog posts, four attendees of the workshop program share insights into their work on data collections and analysis and reflect on the knowledge gained from the interdisciplinary discussions at the Digital Academy. Learn more about the event visiting the Digital Academy Website.

Reflections on uncertainty in a mixed methods study on Heimat, Affect, and Whiteness

by Friederike Windel

As a participant of the Digital Academy in Fall 2023, I was invited to think about uncertainty in my digital research process. Trained in critical social psychology, I would like to take the opportunity to reflect on my dissertation process in relation to digital and social psychological research, emphasizing the need for interdisciplinary dialogue.

Social psychological studies of social media run the risk of reproducing rationalistic, predictability-driven analyses that treat digital data as neutral. Data science reproduces the ideological undercurrent and practices of data science that actively excludes any acknowledgment of a human-made digital space governed by artificial intelligence that “has produced racialized, gendered, and classed models of the self, delivered with imperialist rhetorics of colonization and conquest” (Yarden-Katz 2020, 6).

Part of my dissertation responds to concerns about data science’s reproduction of oppressive practices by making visible examples on German-speaking Twitter of dominant reproductions of Heimat, broadly defined as “home” or “homeland” in German-speaking contexts.

Heimat is a widely used term, both in politics and in everyday language. I was interested in understanding the instances better in which Heimat intersected with racial practices and white privilege in contemporary and historical discussions and constructions of Heimat. My focus was on the affective manifestations of Whiteness and nationalism to challenge and attempt to interrupt rationalist notions that attempt to cleanly categorize constructs like White supremacy or nationalism based on linguistic practices and remove its inherent emotional processes.

Using computational social science (CSS) and discourse analytical methods for part of the dissertation, I analyzed how Heimat is discussed, embodied, and made sense of in affective ways on German-speaking Twitter between January 2018 and July 2022. The dissertation asked how Heimat functions affectively to (re)produce and maintain ideologies like white supremacy and nationalism.

The question of uncertainty and ambiguity was present throughout the process of writing the dissertation.

Uncertainty and the theoretical framework

In its theoretical manifestations, I challenged mainstream definitions of Heimat that claim a neutral, nostalgic, and apolitical relationship with/to the German nation as former interior minister Horst Seehofer aimed for when he added Heimat to the interior ministry’s name in 2018. My interest was in the ways in which this concept has been tightly intertwined with the political, especially, in nationalist interests. Critical theorist Blickle (2002) has shown that discourses around Heimat are revived in the public consciousness periodically when the German nation considers itself unstable and wants to redefine itself. Thus, one of the key societal processes surrounding Heimat is that it becomes more prevalent when there is a feeling of uncertainty about national identity and the desire for more stability. Yet the idea of a stable, coherent nation is a myth that has been reproduced since modern nation-building. To examine the ways in which Heimat is constructed, I looked at everyday practices and constructions of this term.

Another central concept that I used in my dissertation is Whiteness. Similar to Heimat, I was interested in its daily, often implicit, manifestations, where Whiteness is the unnamed and unmarked norm that secures the power of white people (Frankenberg 1988) through material, discursive, and affective processes. I argue that rather than treating Whiteness solely as identity, it is intertwined with the historically constructed, political, cultural, and socio-economic system of white supremacy that subjugates people racialized as not white. This focus on its structural manifestation is as important in digital quantitative and qualitative research. In research on social media platforms like X/Twitter, such focus allows an analysis of questions that are identity-related (like Whiteness) in ways that do not try to “find out” or focus on users’ identity but instead analyze discourses and affect connected to structure.

This desire to “find out” identity in psychological research comes from the dominant perception that experimental research is the gold standard of psychological study. Often, identity categories are portrayed in singular ways or attempted in an intersectional way, added onto each other. These kinds of studies compare categories, estimate effects, or analyze relationships of variables. While such experimental research has an important place for some psychological questions, it risks reproducing ideas of cleanliness of lived experiences and minimizes the influence and impact of structural factors. Furthermore, it reinforces social categories that were artificially created. Thus, psychological research runs the same risk as data science to not tolerate the uncertainty and ambiguity of data. Putting a structural view of Whiteness in the forefront of the analysis thus allows for a better understanding of its connection to everyday processes, as well as for understanding the impact that AI systems have by harming “vulnerable groups and propagate(ing) societal biases” (Jo & Gebru 2020, 306).

The making of Whiteness is a messy process that occurs in practice, in language, in the body, and in affect. Following social psychologist Derek Hook’s theory of the affective technology of Whiteness (2005), I examined the affective formations of Heimat via “ardent forms of belonging which work to assert certain exclusionary relations of cultural/national /historical ownership and privilege” (77). This affective focus, applying the Foucauldian concept of technology, allowed for an analysis of the individualized practices linked to macro-political manifestations of the German state.

Uncertainty in methodology

For the tweet analysis chapters of my dissertation, I used a mixed-methods approach combining the computational social science method of Structural Topic Modeling (STM) and the qualitative method of discourse analysis. The broad computational pattern analysis with STM served as a starting point for a thematic overview of topics present in the dataset. Such analysis also provided descriptive data about the tweets, such as the timing of specific topics and of the tweets overall. CSS requires a critical approach to address rationalist data practices. Social and Data scientist Rob Kitchin (2014) has developed such an approach arguing for an epistemology that is reflexive and acknowledges the data’s social location, the researcher’s positionality, and the politics of the social science conducted.

In addition, CSS research requires the application of social theory and deep contextual knowledge. Thus, a “neutral” presentation of data analyses in my STM tweet chapter was not an option and required me to add contextual explanations as well as social theory that addresses the context further. The subsequent discourse analysis of a subset of tweets allowed for a deeper contextualized analysis of Heimat tweets. This approach acknowledges the uncertainty and messiness of digital data. Rather than finding neat categories for each topic estimated by the STM, a critical approach to CSS allows for an analysis rooted in structural contexts and the variety of Twitter users’ responses to it.

Uncertainty and reflexivity

My approach to the dissertation project and my interpretations of each step of the process are rooted in my lived experiences and positionality. Following Black feminists like bell hooks. (2000) and Patricia Hill Collins (2022), I approached the dissertation from an understanding that the project and my work (the personal) are explicitly political, and that the political is personal. Related to the topic of white supremacy, my positionality as a white, able-bodied, middle-class raised, cis-woman body, I reproduce white norms, risk the centering Whiteness, or misrepresent the reality of systemic racism. These are practices that I continuously try to challenge, and part of this work is embracing the discomfort that comes with making mistakes and reproducing harm. Community and accountability play an important role in this work; and it is often in conversation with others–ones that build on imagining other futures and ones that reflect on mistakes, and highlight our embodied experiences–where an opening and continuation of the work of challenging oppressive systems occurs.

Learning together and uncertainty in missing data

In the research process, interdisciplinary conversations are an important part of thinking together, asking questions, and sharing ideas and approaches to questions. The Digital Academy brought together a group of digital researchers using a multiple range of methods. With a lot of historians sharing their work, I got more interested in thinking about the archive and the question of how we deal with data when it is not there. It left me thinking about the importance of archival research and the qualitative work of oral history, especially for those stories that are erased from archives or made invisible by mainstream media. This question is ever so present in our current political and social realities and demonstrates the need for understanding better digital manifestations of uncertainty as well as the ways in which uncertainty can be used to manipulate and censor groups of people.

Biographical note

Friederike Windel (she/her/hers) is an Assistant Professor in Psychology, Health and Gender at the American University of Paris. Friederike’s scholarship examines constructions of Whiteness, national belonging, exclusion, and affect using computational social science and qualitative research. She is committed to interdisciplinary collaborations and learning about public scholarship.

https://www.aup.edu/profile/fwindel

References

Blickle, Peter. Heimat: A Critical Theory of the German Idea of Homeland. Rochester: Camden House, 2002.

Collins, Patricia Hill. Black Feminist Thought: Knowledge, Consciousness, and the Politics of Empowerment. New York: Routledge, 2022.

Frankenberg, Ruth Alice Emma. White Women, Race Matters: The Social Construction of Whiteness. Santa Cruz: University of California, 1988.

Hook, Derek. "Affecting whiteness: racism as technology of affect." International journal of critical psychology 16 (2005): 74–99.

hooks, bell. Feminist theory: From margin to center. Pluto Press, 2000.

Jo, Eun Seo, and Timnit Gebru. "Lessons from archives: Strategies for collecting sociocultural data in machine learning." Proceedings of the 2020 conference on fairness, accountability, and transparency (January 2020): 306–316. https://doi.org/10.1145/3351095.3372829.

Katz, Yarden. Artificial Whiteness: Politics and Ideology in Artificial Intelligence. New York: Columbia University Press, 2020. https://doi.org/10.7312/katz19490.

Kitchin, Rob. "Big Data, new epistemologies and paradigm shifts." Big data & society 1.1 (April 2014). https://doi.org/10.1177/2053951714528481.

Digital Academy 2023: Dealing with Doubts: Site Georeferencing in Archaeology and in the Geosciences

Center for Uncertainty Studies Blog — Fri, 31 May 2024 15:56:42 +0200

Dealing with doubts: Site Georeferencing in Archaeology and in the Geosciences

by Florian Thiery

Introduction

Archaeological research must handle doubts – in the context of NFDI4Objects¹ often named using the umbrella terms “fuzziness and wobbliness” ((Thiery et al. 2021)) – such as vagueness, uncertainty, and ambiguities in data modelling. Here, “vagueness” is a measure of the precision of a statement; a vague statement is therefore only true to a certain degree. In the case of “uncertainty,” it is completely unknown whether the statement made is true at all (Unold, Thiery, and Mees 2019; Thiery and Mees 2023). This occurs especially in georeferencing. However, creating reproducible and comprehensible data for reuse while guaranteeing data quality in archaeological data involves disclosing doubts and ambiguities (Thiery and Mees 2023). This is also important for data FAIRification, (process to make data Findable, Accessible, Interoperable, Reusable)² addressed in the NFDI. Especially vagueness and uncertainty must be modelled to work with geodata. However, for linking data and FAIRification graph-based modelling as Linked Open Data (LOD) proposed by (Berners-Lee 2006), cf. (Schmidt, Thiery, and Trognitz 2022)) is the technique of choice. Due to the huge variety of research domains, an interdisciplinary, commonly understandable framework for modelling of uncertainties and vagueness in research data is extremely challenging. We will present three data-driven interdisciplinary use cases of dealing with and modelling vague and uncertain geo-references (findspots) based on literature as LOD from the archaeological and geosciences domain.

Modelling Approaches

If we are dealing with doubts such as uncertainties, vagueness and ambiguities, one of the main goals could be to publish and model the following information:

describe where the geoinformation comes from
describe the method of how the coordinate was created
describe the uncertainty issue(s)
use references into the Semantic Web

To do so, I will now discuss the most applied ontologies and modelling approaches done so far in the project. This starts with modelling in Wikidata, then using the Fuzzy Spatial Locations Ontology (FSLO) to create Linked Open Data, followed by modelling in Wikibase instances (which is the base software framework of Wikidata as well).

I. Wikidata

The Wikidata example will be practically explained using the examples of Irish Ogham stones. Ogham stones are monoliths with the early medieval primitive Irish Ogham script, mostly erected on the island of Ireland and in the western part of Britain between the 4th and 9th centuries. Most stones are no longer at the original site, which is important for cartographic recording and makes it more difficult to determine their original function (Macalister 1945).

Figure 1: Florian Thiery and his dad in front of two Ogham stones (left: CIIC 187, the Kilmalkedar Ogham Stone, OSM node 9110402648³; middle/right: CIIC 178, the Coumeenoole Ogham Stone at Dunmore Head, the most western point of Ireland, OSM node 5145413640⁴ on the Dingle Peninsula on an expedition trip. Florian Thiery & Peter Thiery CC BY 4.0.

Entities in Wikidata (here, an Ogham Stone) may have coordinates (P625). These coordinates contain uncertainty and reference information, which can be modelled using Wikidata qualifiers and references. In the case of Ogham Stones, which have been visited in on-site surveys, this can be done as follows (Figure 2):

qualifier: P1180 sourcing circumstances
qualifier: P248 stated in
qualifier: P276 location
qualifier: P459 determination method
qualifier: P2868 subject has role
reference: P11693 OpenStreetMap node ID

Ogham Stones, which are only available in literature and/or online databases such as “Ogham in 3D” or CISP (Figure 3):

qualifier: P248 stated in
qualifier: P3831 object has role
qualifier: P459 determination method
qualifier: P2868 subject has role
reference: P854 reference URL

Figure 2: Ex-situ Ogham Stone with on-site survey. Modelling in Wikidata. Florian Thiery CC BY 4.0.

Figure 3: Ogham Stone: only Literature as Source. Modelling in Wikidata. Florian Thiery CC BY 4.0.

II. Fuzzy Spatial Locations Ontology (FSLO)

The Fuzzy Spatial Locations Ontology is based on famous and well-known ontology in the field such as PROV-O (Provenance Ontology)⁵, SKOS (Simple Knowledge Organization System)⁶ and GeoSPARQL (Standard by the Open Geospatial Concortium (OGC)⁷. It follows the PROV-O concept of Entity, Activity and Agent (Figure 4). In the case of this ontology, Sites (entity) have a Geometry and were created using a Method (activity) by a Person (agent), c.f. Figure 5. The site and geometry include two properties to describe certainty: fsl:certaintyDesc and fsl:certaintyLevel. The resulting Linked Open Data as RDF can be transformed into human-readable HTML files using the SPARQL Unicorn Ontology Documentation research tool (Homburg and Thiery 2024).

Figure 4: Schematic view of the Fuzzy Spatial Locations Ontology based on the PROV-O ontology. Florian Thiery CC BY 4.0.

Figure 5: Schematic view of the Fuzzy Spatial Locations Ontology and its relations, based on the PROV-O ontology. Florian Thiery CC BY 4.0.

III. Wikibase

The Wikibase modelling (Figure 6) is related to the Wikidata modelling approach. Here, a site also has a lat/lon coordinate, which has qualifiers to describe it further:

has certainty level
certainty description
method used
acting person
has source
has source subtype
method description

Figure 6: Wikibase Data Model of a site. Florian Thiery CC BY 4.0.

Use Cases

The aforementioned method can be applied to three use cases from the archaeological and geosciences domain: Ogham Stones in Ireland, Silver Coinage in Croton, and the Eruption of the Campanian Ignimbrite.

I. Irish Ogham Stones (Wikidata)

Ogham Stones sites (as mentioned before) are mentioned in several catalogues such as books, online databases, e.g., the CISP project or online repositories, e.g., Ogham in 3D. These sources give information in different granularities: townlands, descriptions, and coordinates in WGS84/GPS or Irish GRID references. This can be modelled, e.g., in Wikidata using CIIC 242, known as CISP PARAR/1, SMR No. KE066-016005- (Figure 7).

Figure 7: Ogham Stone CIIC 242 (SMR No. KE066-016005-) at the Aghadoe Church Ruin (SMR No. KE066-016001-) near 52,0767683, -9,5543233. Images by Florian Thiery, right-bottom: Map by © Open Street Map Contributors under Open Data Commons Open Database License (ODbL) by the OpenStreetMap Foundation (OSMF), via https://www.openstreetmap.org/node/10040757680 by fthierygeo.Florian Thiery, CC BY 4.0, via Wikimedia Commons.

This CIIC 242 Ogham Stone has the Wikidata Q identifier Q106680977. Geospatial information can be modelled for the on-site survey (in Open Street Map) and the Ogham in 3D database entry (Figure 8). The Open Street Map (OSM) node 10040757680⁸ has the following information: source: survey by user @fthierygeo and the location 52,0767683, -9,5543233. This node is spatially related to the Aghadoe Church (way/378015315; Q30247109) in the Irish County Kerry. On top of that, the Ogham in 3D ID 242._Parkavonear⁹ stated: “GPS coordinates -9.554751, 52.076337.”

Figure 8: Coordinate locations (P625) of Ogham Stone CIIC 242 in Wikidata (Q106680977). Public Domain.

II. Campanian Ignimbrite (CI) ash airfall (LOD / FSLO)

The Campanian Ignimbrite (CI) ash airfall in Central Europe was derived from the Campania region in Italy and a late Pleistocene volcanic event (De Vivo et al. 2001; Barberi, F. et al. 1978). Around 40,000 yr b2k, the largest eruption of the CI took place in the Phlegraean Fields. Based on this, massive glass deposits from the CI-eruption covered large parts of the eastern European continent (Figure 9); volcanic material of the CI is often found in isolated watersheds and valleys. These findspots are recorded in several papers, e.g. precise coordinates, references to cities or references to regions.

Figure 9: Sample of Campanian Ignimbrite Findspots based on several literature. Florian Thiery and Fiona Schenk, CC BY 4.0.

One Findspot is Urluia in Romania. Spatial information about this findspot is mentioned in some papers, such as Fitzsimmons and Hambach (2014, 76), “[...] the site of Urluia Quarry on the Dobrogea loess plateau [...], some 15 km south of the Danube River.” Or Pötter et al. (2021, 5), “The Urluia [...] is located in an abandoned limestone quarry on the limestone plateau of the Dobrogea [...]”. Investigations on OSM showed that this limestone quarry is described using way/84975654, resulting in a coordinate POINT(27.9021 44.0947). This can be modelled using the Fuzzy Spatial Locations Ontology as Linked Open Data in cisite_52.

Another example is a findspot at Crvena Stiljena Cave (Montenegro), mentioned in (Morley and Woodward 2011). This findspot is also an archaeological site described in OSM node/10879170567 and several papers, such as (Morin and Soulier 2017), resulting in the LOD cisite_51.

III. Silver Coinage of Croton (Wikibase)

Research on the silver coinage of Croton, an Achaean colony in southern Italy from the 6th - 3rd century BC, e.g. examines the material under numismatic, historical, and archaeological methods to reconstruct the monetary history of the ancient city. The geo-locations of the finds are almost all not exactly georeferenced (e.g., by use of the documentation of excavation reports). The determined locations are derived from literature and have varying degrees of precision regarding the location of the find spot. An example of the uncertainy information of a find at Esaro River in 1967 shows the challenges:

"There is only the information that the hoard was found on the river Esaro, without specifying exactly where on the river course. In the literature, there is information that it was found "in Crotone", by which the city rather than the province is meant; otherwise, it would have been formulated differently. The urban area of the river course ends shortly after the "Strada Statale 106 Jonica". The most likely location for the find is along the course of the river between the section just before the motorway and the mouth of the river in the sea. Since the find was made in 1967, the river's course at that time probably corresponds fairly closely to the course of the river today. However, both the latter and the presumed find section are conjectures."

This information is modelled using the Wikibase approach as entity Q67 and can be shown in a wikibase.cloud instance for fuzzy-sl at https://fuzzy-sl.wikibase.cloud/entity/Q67.

Digital Academy

The Digital Academy was a nice platform to discuss several approaches to modelling uncertain information in research data. Especially, the totally different views on modelling approaches from various fields, such as the social sciences or natural sciences had let me think about how to better express the ideas to researchers and other people. In this case particularly, the visual representation of the underlying graph data (maybe using the SPARQL Unicorn framework) is the key and must be put more in the focus. The remaining challenges are standardisation of these approaches. Applying the FAIR principles and re-using them needs standardisation by the community. This can be done by the NFDI and especially NFDI4Objects, which deals with uncertainty modelling using semantic modelling approaches.

Biographical note

Florian Thiery is a trained Research Software Engineer (RSE) and Ontology engineer with a master's in Geoinformatics and Surveying from the Fachhochschule Mainz. He currently works at the Leibniz-Zentrum für Archäologie LEIZA as RSE in the NFDI4Objects consortium (Task Area 2: Collecting). More information can be found at http://fthiery.de.

Footnotes

1. https://www.nfdi4objects.net.

2. Cf. https://www.go-fair.org/fair-principles/.

3. https://www.openstreetmap.org/node/9110402648.

4. https://www.openstreetmap.org/node/5145413640.

5. http://www.w3.org/TR/prov-o/.

6. https://www.w3.org/TR/skos-primer/.

7. https://www.ogc.org/standard/geosparql/.

8. https://www.openstreetmap.org/node/10040757680.

9. https://ogham.celt.dias.ie/242._Parkavonear

References

Barberi, F., Innocenti, F., Lirer, L., Munno, R., Pescatore, T., and Santacroce, R. 1978. ‘The Campanian Ignimbrite: A Major Prehistoric Eruption in the Neapolitan Area (Italy)’. Bulletin of Volcanology 41.

Berners-Lee, Tim. 2006. ‘Linked Data - Design Issues’. 27 July 2006. https://www.w3.org/DesignIssues/LinkedData.html. Accessed March 15, 2024.
De Vivo, B., G. Rolandi, P. B. Gans, A. Calvert, W. A. Bohrson, F. J. Spera, and H. E. Belkin. 2001. ‘New Constraints on the Pyroclastic Eruptive History of the Campanian Volcanic Plain (Italy)’. Mineralogy and Petrology 73 (1–3): 47–65. https://doi.org/10.1007/s007100170010.

Fitzsimmons, Kathryn E., and Ulrich Hambach. 2014. ‘Loess Accumulation during the Last Glacial Maximum: Evidence from Urluia, Southeastern Romania’. Quaternary International 334–335 (June): 74–85. https://doi.org/10.1016/j.quaint.2013.08.005.

Homburg, Timo, and Florian Thiery. 2024. ‘Sparqlunicorn/sparqlunicornGoesGIS-Ontdoc: Version 0.17’. Squirrel Papers 6 (2): #2. https://doi.org/10.5281/zenodo.8190763.

Macalister, Robert Alexander Stewart. Corpus Inscriptionum Insularum Celticarum. Vol. I. Dublin: Stationery Office, 1945.

Morin, Eugene, and Marie-Cécile Soulier. ‘The Paleolithic Faunal Remains from Crvena Stijena’. In Crvena Stijena in Cultural and Ecological Context. Multidisciplinary Archaeological Research in Montenegro, edited by Robert Whallon, 266–294. Podgorica: Montenegrin Academy of Sciences and Arts, 2017.

Morley, Mike W., and Jamie C. Woodward. 2011. ‘The Campanian Ignimbrite (Y5) Tephra at Crvena Stijena Rockshelter, Montenegro’. Quaternary Research 75 (3): 683–96. https://doi.org/10.1016/j.yqres.2011.02.005.

Pötter, Stephan, Daniel Veres, Yunus Baykal, Janina J. Nett, Philipp Schulte, Ulrich Hambach, and Frank Lehmkuhl. 2021. ‘Disentangling Sedimentary Pathways for the Pleniglacial Lower Danube Loess Based on Geochemical Signatures’. Frontiers in Earth Science 9 (April): 600010. https://doi.org/10.3389/feart.2021.600010.

Schmidt, Sophie C., Florian Thiery, and Martina Trognitz. 2022. ‘Practices of Linked Open Data in Archaeology and Their Realisation in Wikidata’. Digital 2 (3): 333–64. https://doi.org/10.3390/digital2030019.

Thiery, Florian, and Allard Mees. 2023. ‘Taming Ambiguity - Dealing with Doubts in Archaeological Datasets Using LOD’. CAA 2018: Human History and Digital Future, October. https://doi.org/10.15496/PUBLIKATION-87762.

Thiery, Florian, Allard Mees, Karsten Tolle, and David Wigg-Wolf. 2021. ‘TRAIL2.2: Evaluation of Fuzziness and Wobbliness in Numismatics and Ceramology’. NFDI4Objects TRAILS 2021 (October): No. 2.2. https://doi.org/10.5281/zenodo.5654897.

Unold, Martin, Florian Thiery, and Allard Mees. 2019. ‘Academic Meta Tool. Ein Web-Tool Zur Modellierung von Vagheit’. ZfdG - Zeitschrift Für Digitale Geisteswissenschaften Die Modellierung des Zweifels – Schlüsselideen und-konzepte zur graphbasierten Modellierung von Unsicherheiten. (Sonderband 4). https://doi.org/10.17175/SB004_004.

Digital Academy 2023: Exploring Uncertainty in Toponyms within the British Colonial Corpus

Center for Uncertainty Studies Blog — Thu, 2 May 2024 17:55:22 +0200

Exploring Uncertainty in Toponyms within the British Colonial Corpus

by Shanmugapriya T

My research project aims to extract toponyms from the British India colonial corpus to create a historical gazetteer. The primary challenge in this work revolves around the toponyms themselves, as they exhibit a high degree of fuzziness and inconsistency, particularly in their spellings. Historically, mapping, documenting, and surveying have been recognized as essential tools employed by colonial powers to demarcate, expand, and exert control over their colonial subjects. These activities enabled the colonial administration to establish governance over land and streamline revenue collection during the British colonial period. As time progressed, surveys expanded beyond their initial military and geographical purposes, evolving into comprehensive sources of information encompassing geography, political economy, and natural history. The British colonial India corpus is, therefore, intricate, marked by non-standard formatting, and plagued by inconsistencies in the spelling of Indian toponyms. This intricacy adds an extra layer of complexity to the task of extracting and organizing these toponyms for the creation of a historical gazetteer. The recognition of these challenges underscores the importance of using advanced techniques and tools to handle the uncertainty inherent in this historical data.

Digital Humanities methods and tools

Dealing with fuzzy toponyms requires the application of specific and advanced techniques. In this context, I utilize digital humanities methods and tools to identify and extract these toponyms from the British India colonial corpus. Indian toponyms in the British colonial corpus often exhibit various spellings, such as "Noil", "Noyal", "Noyyal", "Bawani", "Bhawani" and "Bowani," representing different variations of river and place names in the Southern region of India. To address this challenge, I conducted an exploration of the corpus. My approach involved leveraging an English word database, employing regular expressions, using natural language processing module Spacy for customized entities, and utilizing other relevant Python libraries to extract transliterated words from the corpus. Additionally, I developed a user interface using programming languages HTML, CSS and JavaScript. I used an open access database MySQL to store the data and PHP for interactive and management of the data. Finally, I employed Geographic Information System (GIS) tool ArcGIS to filter, map, and tag the toponyms and other entities within the dataset. While these initial experiments contributed to theoretical considerations and raised awareness of the complexities inherent in studying the British colonial corpus, the employed method did not entirely resolve the challenge of extracting toponyms. It also inadvertently filtered out misspelled and non-contemporary English words, along with the targeted toponyms.

The new method I propose involves three distinct stages. The first stage centers on the identification of entities using advanced natural language processing module BERT Named Entity Recognition (NER) (Devlin et al. 2018) to create a trained dataset on place names. This NER system is instrumental in locating hidden toponyms and learning from contextual information. The second stage is dedicated to the extraction of fuzzy toponyms, for which I employ advanced natural language processing module DeezyMatch (Hosseini et al. 2020). DeezyMatch is specifically designed for fuzzy string matching and toponym extraction. To generate the training dataset for string pairs, I also collect alternate names of places in South India. By learning similar transformations as those present in the training set, DeezyMatch should be capable of applying this knowledge to unseen variations of toponyms. Subsequently, I use the cleaned dataset to determine optimal hyperparameters for specific scenarios, such as finding the ideal thresholds for matching. In the final stage, I create a database for the historical gazetteer and integrate it with the World Historical Gazetteer. This integration is significant as it offers a wide range of content and services that empower global historians, their students, and the general public to engage in spatial and temporal analysis and visualization within a data-rich environment, spanning global and trans-regional scales (“Introducing the World Historical Gazetteer”). This enhances the accessibility and utility of the historical toponym data for a broad audience.

Main challenges

The first and foremost challenge is the absence of a trained dataset of Indian place names. I need to focus on creating a trained dataset using Named Entity Recognition and other external open-access resources, such as Wikipedia. The second challenge pertains to the advanced programming techniques that I am experimenting with. The initial experiment with BERT NER for identifying toponym entities demonstrates that the algorithm performs well compared to other NER libraries. However, it also identified a few words that are not toponyms as place names and did not identify the broken toponym words as place names. Therefore, the extracted place name entities will require manual verification to confirm their accuracy. I anticipate encountering additional challenges when I begin exploring DeezyMatch, as I am currently in the initial stages of my research.

Digital Academy workshop on uncertainty

The Digital Academy workshop presented a fantastic opportunity for scholars like myself to convene and discuss a wide array of challenges, approaches, methods, and tools for addressing uncertainty. The inclusion of experts in the field of uncertainty was a valuable aspect of this workshop, enabling attendees to solicit advice and feedback on the challenges they face in their research. Although I was not able to attend the entire workshop, the workshop's theme serves as a motivating factor for me to persist in my research endeavors despite the numerous challenges I've encountered. I believe that ongoing discussions and collaboration within the academic community will be instrumental in finding effective solutions to these challenges and further advancing the field.

Questions remain open

The open questions revolve around the ideal size of the corpus required for applying the aforementioned advanced techniques and the expected effectiveness of the trained dataset. However, I am hopeful that I will be able to find answers to these questions in the near future.

References

World Historical Gazetteer. “Introducing the World Historical Gazetteer.” Accessed October 10, 2023. https://whgazetteer.org/about/.

Devlin, Jacob, Ming-Wei Chang, Kenton Lee and Kristina Toutanova. “BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding.” North American Chapter of the Association for Computational Linguistics (2019). Accessed October 5, 2023. https://arxiv.org/pdf/1810.04805v2.

Hosseini, Kasra, Federico Nanni, and Mariona Coll Ardanuy. “DeezyMatch: A Flexible Deep Learning Approach to Fuzzy String Matching.” Paper presented at the Empirical Methods in Natural Language Processing: System Demonstrations, Online, October 2020. https://aclanthology.org/2020.emnlp-demos.9. Accessed October 5, 2023.

Biographical note

Shanmugapriya T is an Assistant Professor in the Department of Humanities and Social Sciences at the Indian Institute of Technology (Indian School of Mines) Dhanbad. She was a Digital Humanities Postdoctoral Scholar in the Department of Historical and Cultural Studies (HCS) at the University of Toronto Scarborough. Her expertise centers around the development and application of digital humanities methods and tools for historical and literary research in South Asia, particularly within the realms of colonial and postcolonial studies. She has a specific focus on areas such as text mining, digital mapping, and the creation of digital creative visualizations.

Visit the personal website: https://www.shanmugapriya.com/

Digital Academy 2023: Catrina Langenegger about Swiss Military Refugee Camps

Center for Uncertainty Studies Blog — Fri, 5 Apr 2024 11:49:26 +0200

Historical Map of Switzerland.

Swiss military refugee camps

by Catrina Langenegger

In my research project I examine the Swiss policy of asylum and the military camps for refugees during the Second World War. In this blog post, I thereby focus on the data I collected on these refugee camps and the questions of uncertainty within my work with the data. I encountered uncertainty primarily in the areas of incomplete data, the standardisation process and different data qualities. I will first give a short introduction to my research topic and will then discuss the sources and data I collected. I will thereafter focus on my work with the data, the challenges I encountered when dealing with uncertainty and the benefits I took away from the Digital Academy.

Refugee aid is a civil task. As I focus on military support, I consequently deal with a temporary, exceptional phenomenon. In Switzerland, first the private refugee aid organisations and then the department of police were responsible for the refugees. From 1940 onwards the department of police opened camps to home the refugees and emigrants who sought protection in Switzerland. In the late summer of 1942 the number of refugees was constantly rising. More and more, the civil administration was overstrained. It could neither provide enough space for housing, nor enough financial support, food and staff. Briefly said, the system of civil refugee camps was in danger to collapse. In this situation, the military was asked to stand in. The army was considered to be the only institution that could acquire enough buildings, recruit enough personal and provide a sufficient system for replenishment.

In September 1942 the first reception camp lead by the military was established in Geneva. The army took over the first care for the refugees with food, clothing and accommodation. From that point of time, a new system of three different camps lead by the military was established, that every refugee hat to go through, before being placed constantly in a refugee camp under civil administration. Collecting camps where placed next to the boarder. Due to concerns for hygiene, the refugees were obliged to spend three weeks in a quarantine camp. After the quarantine, the refugees could theoretically move to civil camps but most of the refugees had to stay in reception camps because there was no space for them under the civil administration. Some of the refugees had to stay only for a few days or weeks, others spent months in reception camps. These military refugee camps are the topic of my research. They operated until after the end of the war.

Serial sources as data

Besides the administrative sources like commands and instructions, protocols of inspections and meetings, and weekly reports from the camps are stored in the Swiss federal archives. These serial sources are the basis of my data analysis. I found them in eleven different archive collections. I extracted the information out of the reports into a database. All in all, I found reports on 168 weeks, from October 1942 to July 1946. Nevertheless, the thereby combined collection contains voids. For at least eleven weeks no reports were to be found. It is at least eleven because the first report dates on the 18th of October 1942. However, first camps were opened in September 1942. I am not aware of earlier reports as I could not find any. But it is also possible that the standardised reporting started only in the middle of October. The voids are one aspect of uncertainty I will focus on in this blog post. I aim at being transparent about the gaps and make them visible at all stages of processing.

During the process of data cleaning, I decided to work only with data that refers to one or more refugee in a camp. Data with no refugees or camps that were emptied and only on reserve are therefore not included in the dataset. All in all, I have a dataset with more than 6’000 observations on refugees in the camps. These observations do not only show how many refugees were housed, but also which type of refugees (civilian, military) they were and which type of camp (quarantine, collection, or reception camp) it was. Reflecting on these categories is part of my data critique and leads as well into the field of uncertainty.

The next step was data cleaning and standardisation. I corrected obvious typing errors in the process of data extraction to reduce the number of variables. Then I standardised the camp names. As a subject librarian, dealing with data and meta-data as well as standardising it is part of my daily task. Here are some examples for standardisation with changing names: the camp name “Grand Verger” refers to the same camp as “Signal”. Similarly, the names “Geisshubel” and “Rothrist” refer to the same camp. I put a lot of effort into the standardisation. In the end I found 221 camps. Since one aim of my research project is to depict and analyse the refugee camp system over time, it was important to have a data set as clean and reliable as possible as a basis for the analysis. The various standardisation steps were important for data quality, as the quality of the entire analysis depends on it.

Handling data and uncertainty

To take a step further and to focus on questions about living in the camps during the analysis, I enriched my data with information about the building type and the exact georeference. My approach to deal with the uncertainty I encountered when collecting geodata for every camp to analyse and visualize it in a geographic information system (GIS) to show the geographical distribution, was triangulation by other source types. Sources that contained the necessary information were reports, protocols, autobiographies etc. I also used historical maps provided by swisstopo¹, to localise the camps. In many cases the information was good: “factory building 500 metres outside the village” or “Hotel up on a hill between this village and the other”. I could then add the exact geodata. For other camps, the information was not as precise as I had hoped for, and I had only the name of the village. In other cases – most of them were hospitals, prisons, or camps that were only open for a short time. But the location was always within the borders of the territorial district. So I made a sound decision for these camps. For one entity without any information, not even the district, I decided to not georeference it at all.

As I am working as a librarian, I am used to the convention of coding the quality of the metadata. In a library catalogue you can check the level of cataloguing, whether the book was catalogued by a librarian or a machine for example. Having varying qualities of data in my set, I aimed at qualifying it. I therefore went for three different categories: A B and C to make a statement on the accuracy of my data. If someone wants to use my data later, the uncertainty is made transparent through this code. A stands for the best quality, i.e. information about the address at the level of the building. B stands for medium quality; the information is correct at the village or town level. C stands for the most uncertain category, the information is provided within the territorial district and is based on variant indications.

I now come back to the missing reports mentioned above. My goal is to be transparent about this gap. However, making this gap visible in statistics and visualisations is one of the greatest challenges when dealing with uncertainty. Statistics and visualisations are positivistic: they only show what is there. In the first statistics, the gaps weren’t visible. I therefore made artificial observations in my dataset with a zero as value to mark the gaps. In other words, I made the missing weekly reports visible by creating an observation for each of these dates. I have labelled these artificial observations as such. My data model now provides a field to mark whether there is a report for the week or not. Nevertheless, it’s almost impossible to visualise the weeks without information. Although I have made artificial entries in my dataset, these are not displayed in the visualizations because they do not contain a value.

fig. 1: Timeline with missing data

fig. 2: Auto-corrected timeline

The software I use calculates out all uncertain data and provides the average. I found a way to work around this by only using the edit mode, even for my visualisations because in the viewing mode, the observations inserted by me to show the uncertainty will be removed. In both examples, I was able to incorporate the uncertainty into the data via a categorisation in my data model. In this way, I also hope that my data can be better reused, as it makes transparent statements about its own quality.

The workshop of the Digital Academy 2023 gave me the impetus to take a closer look at the subject of insecurity. The opportunity to exchange ideas with other researchers was very enriching. I was also able to present how I deal with uncertainty and develop an even clearer definition of my categories and my approach based on the discussions and comments in the workshop.

Biographical note

Catrina Langenegger recently submitted her PhD thesis on refugee camps under military control in Switzerland during the Second World War. She conducts her research at the Centre for Jewish Studies at the University of Basel. As a historian with a focus on digital humanities she exercises her passion for data also in her role as subject librarian with a background in library and information sciences.

References:

1. Cf. Karten der Schweiz - Schweizerische Eidgenossenschaft - map.geo.admin.ch: https://map.geo.admin.ch/?topic=swisstopo&lang=de&bgLayer=ch.swisstopo.pixelkarte-farbe&catalogNodes=1392&layers=ch.swisstopo.zeitreihen&time=1864&layers_timestamp=18641231.

Meet ... Jens Zinn

Center for Uncertainty Studies Blog — Wed, 6 Dec 2023 09:15:37 +0100

Jens Zinn is Tr Ashworth Associate Professor in Sociology Social and Political Sciences at The University of Melbourne and CeUS Member.

What connects you to Bielefeld University?

I am connected to Bielefeld University personally and professionally. After my undergraduate studies in Saabrücken I was attracted by the large and only Faculty of Sociology in Europe at Bielefeld University which offered a large variety of approaches taught by outstanding sociology scholars. It was also a formative experience since I learned about concepts such as ‘time’ and ‘risk’ which became influential in sociological debate (Beck 1986, 1988; Luhmann 1985, 1991; Douglas & Wildavsky 1982). Amongst the many scholars in particular the analytical sharpness of Niklas Luhmann and Franz Xaver Kaufmann but also the historical work of Reinhard Koselleck influenced my work and approach to risk and uncertainty as analytical concepts as well as discourse semantic changes.

I am therefore still connected to the scholarship in systems theory and the Institute for World Society Studies as well as Historical Semantics and the corpus/computational analysis of social change (compare SFB 1288).

Indeed, having lived and worked at Bielefeld University I am also emotionally attached to the central university building (I consider as “the starship”) providing everything what is needed to focus on research. As a research assistant, from my office I had a good view on the Centre of Interdisciplinary Research (ZiF), which was already in the early days an indication for the innovative interdisciplinary research culture at Bielefeld University. With the University being placed close to the “Teuto” (Teutoburger Wald) I still enjoy walking through the woods whenever I find the time when visiting.

What role does Uncertainty play in your research?

Uncertainty is a key concept in my research. When I initiated two research networks on risk studies within the Europeans Sociological Association (2005) and the International Sociological Association (2006) I was keen to find the key concepts which could hold together the complex scholarship on risk studies and would characterise a broader sociological rather than a psychological, economical or technological approach to the future. At this point the Sociology of Risk and Uncertainty (SoRU) was born to see risk in the context of uncertainty, and uncertainty in the context of its social relevance when something of value is at stake (this includes possible harm as well as gains but the recognition of the relevance of the unknown for the presence). In this way risky uncertainty characterises decision making situations I am interested in. These contrast with people following worn-out paths of routines without further consideration.

It is here where my recent work on everyday life engagement with risky uncertainty connects with uncertainty studies. In the social realm the modernisation process contributed to a significant shift in the ways how uncertainty is understood and managed. A key element has been the development of calculative technologies and most recently the advancement in computer technology and advancing social digitisation. At the same time ‘hope’, ‘faith’ and ‘ideology’ remain powerful resources of social enchantment which seem also necessary for managing risky uncertainty. The comparatively abstract forms of reasoning related to social rationalisation and enchantment are not sufficient to understand people’s engagement with risky uncertainty in everyday life. Here the subjectivation of detached forms of knowledge are required to understand lifeworld forms of reasoning represented by concepts such as ‘trust’, ‘intuition’ and ‘emotions’ (Schulz & Zinn 2023).

What would you like to accomplish in a Center for Uncertainty Studies?

The Centre of Uncertainty studies is an exciting hub which opens opportunities for interdisciplinary collaboration and innovative conceptual advancement. There are three areas of research I expect to advance in the Centre for Uncertainty Studies.

(1) In many ways the different disciplines involved in CeUS represent different understandings of uncertainty which are influential in public debates. I am interested in the imaginaries and research practices through which my colleagues construct uncertainty as a research object as well as a research reality to be managed. On this basis I would like to further develop an outline of the sociology of uncertainty and risk, which helps to specify and understand how social forces combine or amalgamate in the social navigation of uncertainty.

(2) In a more concrete conceptual enterprise, I want to further develop a phenomenology of uncertainty and risk, which is capable of making sense of the processes of the individual and institutional engagement with risky uncertainties. This would follow developments in social science disciplines which not only study modes of engagement with uncertainty such as trust, intuition, emotions, and hope but how such modes inform the research process itself.

(3) The broad methodological expertise within CeUS allows developing digital resources and methods to analyse the societal understanding and responses to risky uncertainties such as pandemics, climate change related new social challenges (e.g., heat waves and other weather events). I would like to advance the collaboration between different disciplines such as linguistics, sociology, history and digital humanities more broadly and mathematically trained modelers to develop powerful research instruments (conceptually and technically) to better understand historical developments as well as the meaning and effects of increasing societal digitisation.

To what extent is interdisciplinarity important in your work?

The social management of risk and uncertainty relies on interdisciplinary and transdisciplinary approaches. These support collaborative learning which is crucial for producing good and socially acceptable outcomes. In this context my sociological approach to uncertainty also profits from connecting to other research such as in psychology, media studies, health studies, history, linguistics, and philosophy. Insights from risk perception studies as well as decision making research has informed my studies as well as linguistic research instruments for the analysis of discourse semantic changes of risk. Conceptual insights from philosophy are informing my theoretical work on a phenomenology of risk and uncertainty as well as empirical insights from environmental sociology, science and technology studies, disaster research, media studies, health studies and youth studies and many more. Thus, being strongly rooted in sociology my research and theorizing is informed and connects to other disciplines to show its relevance across disciplines as much as getting inspired by related work from different disciplinary perspectives.

The first CeUS conference ("Navigating Uncertainty: Preparing Society for the Future") took place in Bielefeld at the beginning of June - which moments were particularly exciting for you? What do you take away?

The conference became quite exciting when I realised to what extent my own conceptual work on ‘rational’, ‘non-rational’ and ‘in-between’ modes of engaging with risk and uncertainty can connect to the empirical work presented by many of the participants considering trust, emotions, hope and other ways of engaging with uncertainty.

Admittedly, I was not able to connect to every contribution in the same way. However, I was surprised and thrilled by the large number of disciplines I could connect to such as conflict studies and historical studies.

To sum it up: Do you have specific strategies in your personal or professional life to deal with uncertainty?

My approach as well as my professional interest relate to what Greek philosophers might have positioned in the realm of “phronesis”. This seems to me a decent way to engage with risky uncertainties which cannot be easily mastered by the application of technique or differentiated knowledge systems but requires practical wisdom which considers ethical and normative standards as well as different forms of (non-)knowledge in research, professional decision making and the life world.