A (not so) short history of historical network research – part 1

by Cindarella Petz

Buckle up for a dive into the history of historical network research! In this (three-part series) post, I will recount the history of computational thought in historical research, discuss networks for history, and highlight some of the cornerstones of network methodologies in historical studies.¹ Prepare for part 1 on computational thought in the historical studies.

Before we start this quick deep dive (remember, the amount of air needed increases as pressure increases by 1 bar for every 10m going down. So at 20m down, we need three times the amount of air and can stay only 1/3 the time as at surface level!), we have to check our gear.

So I am prefacing this with a couple of remarks:

1. 1. We can define historical network research as consisting of methodologies and approaches from the interdisciplinary field of network research (which itself combines input from sociology, anthropology, political studies, mathematics, and computer science) to study network models of the past. When these are embedded into a larger context of historical research, historical network analysis can help generate greater understanding and new insights into historical objects of research.
   2. Historical network analysis (HNA) and social network research (SNA) use the same set of methods and concepts; this has been remarked, too, by my colleague Demival Vasques Filho in his recent “Introduction to historical (social) network analysis.” But HNA and SNA differ in the sources used: for HNA, it’s usually historical sources in historical research, while SNA more contemporary sources are used in the social studies (such as social media data, or data generated for just that purpose as through interviews or surveys). The different source basis results in (sometimes) different limitations of similar applications of networks within the respective fields, like the fragmentation of sources.
   3. Networks are semantically annotated graphs with a set of links among a set of entities “with the additional property that the characteristics of these linkages as a whole may be used to interpret the social behavior of the persons involved,” as in the commonly used definition by Mitchell (1969, p. 2). Network research focuses on the interdependence of entities (Wasserman and Faust 1994, p. 4), which shape their scope of actions: as such, the position of an entity in a social network is of importance–either for power, performance, or the ‘capital’ of potential social exchange. Freeman (2004, p. 3) characterized (social) network research as (a) the (intuitive) analysis of the social relations of actors, (b) the systematic analysis based on empirical data, (c) their graphical representation, and (d) mathematical, computer-assisted formal model to analyze such data. The requisite of a formalized analysis within network research provides a key point in evaluating historical network research: a metaphorical usage of the network terminology lacks such a systematization. But more on this later on.
   4. I want to stress here that networks are inherently models. As models, networks offer an approximation to and of reality through the (ultimately always) subjective selection of what are supposedly important or representative factors or parameters and their subjective weighting and evaluation, following the classic definition of models by Stachowiak (1973). Thus, modeling is a fundamental epistemic tool: “We model to understand,” paraphrased Le Moigne (2004) Stachowiak’s take (1973, p. 56) that all insights are insights in models or through models (compare also to Piotrowski 2020, p. 10). As approximations to reality, however, any model is subjectively biased. Consequently, any model can be considered only one instance of an “endless number of potential other models” (Rehbein 2020, p. 265), and as such, it offers more of a ‘crutch’ but never the ‘truth’ on the object of research Drucker (2011). This dichotomy of network models as a ‘crutch’ to approximate (an historical) reality and the formalized analysis of such models allow for a heterogeneous span of applications of the network terminology and methodology in historical research.

The history of historical network research, in my opinion, cannot be characterized as a genealogy: usually, historical studies employing a network methodology did not progressively build upon each other (unless within research groups or associations), and reference instead the inaugurating methodological studies of the general (social) network research. Therefore, I will contextualize the history of HNR by its two influential streams of origin: for one, the long history of computational thought and computation-based methods in the historical studies (a much longer history, than many would like to credit, and part 1 of this blog post series); and for two, the developments in the field of social network research (part 2); before I will tackle networks in historical research (part 3).

Computational thought in historical research

Since the early 20^th century, the French Annales school famously heralded the then-new social history to utilize previously ignored types of sources (such as census records) in order to grasp and quantify the experience of the ‘ordinary’ people – in contrast to the study of elites, colloquially known as the history of the ‘great men’ (compare, e.g., to Reynolds 1998, pp. 141-2; Aydelotte 1966, Boonstra et al. [2004] 2006, p. 25; Anderson 2007; Crymble 2021, p. 22). This would evolve, in the wake of computerization, to the new field of quantitative history, also called ‘cliometrics’ or ‘econometrics.’

However, there are also examples of proto-computational studies from the 17^th and 18^th centuries before the professionalization of the historical studies (compare to, e.g., Gordon (1991) 2003; Novick 1988; Crymble 2021, pp. 20–7; Thaller 2017). This post’s entry lithography is an early example of data visualization by Charles Minard, and depicts the losses of the French army on their trek to Russia in 1812-13 over specific geographical locations and temporal events, showing the army’s strength thinning on their way towards Moscow (in red) and on their way back (in black). Coincidentally, it can also be considered a network visualization.

In the 1960s, machine-based computational approaches (which entailed quantitative approaches and statistical analysis) became popular in the humanities, then under the moniker of ‘humanities computing’ or ‘computing in the humanities.’ Foreshadowing future developments here already: Only in the early 2000s, ‘humanities computing’ would grow to become ‘Digital Humanities’ on the onset of mass digitization and the birth of the Internet!

This spread of computational experimenting in the humanities was a result of the technological advances in computing and the steadily increasing availability of (human and nonhuman) computers and statistical software, which allowed to search for patterns and to generate new insights through quantification using machines (Hockey 2004, p. 5; Crymble 2021, p. 24). First computers were human, usually female statisticians (see, for example, the photograph of the astronomer Edward Charles Pickering’s female computers, themselves oftentimes accomplished astronomers); military research brought advances in cutting-edge proto-computers in the 1940s, such as punch-card reading machines, which again were for a great part operated by female workers.

Soon, these made place for room-filling mainframe computers, which continuously shrank in size since the late 1950s, culminating in the development of the portable personal computer in the 1980s (compare to, e.g., Anderson 2007; Crymble 2021, pp. 21—2; Petz 2022, pp. 17—8).). From ‘humanities computing’ soon sub-disciplines emerged: for instance, ‘computer linguistics’ (which subsequently made the jump to an independent research field), or specialized sub-fields such as ‘history and computing’ and the ‘historical information science’ (Boonstra et al. [2004] 2006).

Alongside the advances of computational prowess, and independent of the developments in ‘humanities computing,’ especially a younger generation of historians turned to quantitative history (‘cliometrics’) in its heyday during the 1960s until the late 1970s, as Whales (1991, p. 296) has shown. This was facilitated by the increased access to adequate technical support (Aydelotte 1966; Reynolds 1998, pp. 141–2; Boonstra et al. [2004] 2006; Anderson 2007), as well as by the rising interest in social and economic history in the general historical profession.

Though quantitative methodologies were used in historical research projects and discussed in traditional journals and specialized conferences, and integrated into university curricula (Anderson (2007) noted that statistical methods have been part of 40% of US curricula), they also experienced resistance and backlash against its perceived positivism by the more ‘orthodox’ historical studies, ultimately leading to its estrangement by the mid-1980s following on foot the implications of the post-structuralist cultural turn and changed focus to the social nature of texts of the linguist turn (Graham et al. 2016, p. 23; Petz 2022, p. 19). “Texts replaced tables,” summarized this development Edward L. Ayers (1999), when the popular quantitative history began to shrink to a niche.

From the 1980s on, computers had finally “pervaded every conceivable field of the humanities,” asserted Boonstra et al. ([2004] 2006, p. 13): the personal computer and electronic mail slowly were integrated into the humanities researcher’s life, and working with computers became in general easier and quicker due to advances in quicker file processing (Hockey 2004, pp. 9–10). The advent of the internet heralded a new era, followed on foot with the beginning of mass digitization in the mid-1990s. But, that computers were available and capable of being used for computation-based analysis and text processing didn’t (and still doesn’t) mean that they were actually used in this way by many.

In the upcoming part 2 of this series, I will elaborate on the development and underlying assumptions of general (social) network research, in order to tie this together for historical network research in the third installment of this series!

References

• Anderson, Margo. 2007. “Quantitative History.” In The SAGE Handbook of Social Science Methodology, edited by William Outhwaite and Stephen P. Turner, 248–264. London: Sage. doi:10.4135/ 9781848607958.n14.
• Aydelotte, William O. 1966. “Quantification in History.” The American Historical Review 71 (3): 803–825.
• Ayers, Edward L. 1999. “The Pasts and Futures of Digital History.” VCDH Virginia Center for Digital History (website). http://www.vcdh.virginia.edu/PastsFutures.html.
• Boonstra, Onno, Breure, Leen, and Doorn, Peter. [2004] 2006. “Past, Present and Future of His- torical Information Science.” (Amsterdam).
• Crymble, Adam. 2021. Technology and the Historian. Transformation in the Digital Age. Urbana, Chicago, Springfield: University of Illinois Press.
• Drucker, Johanna. 2011. “Humanities Approaches to Graphical Display.” Digital Humanities Quarterly 5 (1). http://www.digitalhumanities.org/dhq/vol/5/1/000091/000091.html.
• Freeman, Linton C. 2004. The development of social network analysis: a study in the sociology of science. Vancouver, BC: Empirical Press.
• Gordon, Scott. (1991) 2003. The history and philosophy of social science. Taylor & Francis e-Library.
• Graham, Shawn, Milligan, Ian, and Weingart, Scott. 2016. Exploring Big Historical Data. The Historian’s Macroscope. London: Imperial College Press.
• Hockey, Susan. 2004. “The History of Humanities Computing.” In A Companion to Digital Humanities, edited by Susan Schreibman, Ray Siemens, and John Unsworth. Oxford: Blackwell. http://digitalhumanities.org:3030/companion/view?docId=blackwell/9781405103213/9781405103213.xml&chunk.id=ss1-2-1
• Le Moigne, Jean-Louis. 2004. Le Constructivisme. Tome III. Mod ́eliser pour comprendre. Collection Ingénium. L’Harmattan.
• Mitchell, J. Clyde. 1969. “The Concept and Use of Social Networks.” In Social Networks in Urban Situations: Analyses edited by J. Clyde Mitchell, 1–50. Oxford: Published for the Institute for Social Research University of Zambia by Manchester University Press.
• Novick, Peter. 1988. That Noble Dream. The `Objectivity Question’ and the American Historical Profession. Cambridge University Press.
• Petz, Cindarella. 2022. “On Combining Network Research and Computational Methods on Historical Research Questions and its Implications for the Digital Humanities.” Dissertation at the TUM School of Social Sciences and Technology, Technical University Munich. http://mediatum.ub.tum.de/node?id=1624881.
• Piotrowski, Michael. 2020. “Ain’t No Way Around It: Why We Need to Be Clear About What We Mean by ”Digital Humanities”.” Preprint submitted for the Pproceedings of the Symposium “Wozu Digitale Geiteswissenschaft? Innovationen, Revisionen, Binnenkonflikte,” held at Leuphana University Lüneburg, Nov. 20 – 22, 2019. doi:10.31235/osf.io/d2kb6.
• Rehbein, Malte. 2020. “Historical Network Research, Digital History, and Digital Humanities.” In The Power of Networks. Prospects of Historical Network Research, edited by Marten Du ̈ring, Florian Kerschbaumer, Linda von Keyserlingk-Rehbein, and Martin Stark, 253–279. London New York: Routledge.
• Reynolds, John F. 1998. “Do Historians Count Anymore? The Status of Quantitative Methods in History, 1975–1995.” Historical Methods: A Journal of Quantitative and Interdisciplinary History 31 (4): 141–148.
• Stachowiak, Herbert. 1973. Allgemeine Modelltheorie. Wien, New York: Springer Verlag. https://archive.org/details/Stachowiak1973AllgemeineModelltheorie.
• Thaller, Manfred. 2017. “Geschichte der Digital Humanities.” In Digital Humanities: Eine Einführung, edited by Fotis Jannidis, Hubertus Kohle, and Malte Rehbein, 3–12. Stuttgart: J. B. Metzler.
• Wasserman, Stanley and Faust, Katherine. 1994. Social Network Analysis: Methods and Applications. 1st ed. Structural Analysis in the Social Sciences, 8. New York: Cambridge University Press.
• Whales, Robert. 1991. “A Quantitative History of the Journal of Economic History and the Cliometric Revolution.” The Journal of Economic History 51 (2): 289–301.

---

Image credits: Featured image: Charles-Joseph Minard: Carte figurative des pertes successives en hommes de l’armée française dans la Campagne de Russie 1812-1813. Régnier & Dourdet 1869. https://gallica.bnf.fr/ark:/12148/btv1b52504201x/f1.item.zoom#.

Fig 3.: Betty Jennings (left) and Frances Bilas (right), two of six women chosen to work as the main programmers for the ENIAC, arranging the program settings on the master programmer, 1947. Archival information: 1985231_001_001_005, Box 1, Folder 1 , Sperry Corporation, UNIVAC Division photographs and audiovisual materials (Accession 1985.261), Audiovisual Collections and Digital Initiatives Department, Hagley Museum and Library, Wilmington, DE 19807. https://digital.hagley.org/1985231_001_001_005. Listed as Public Domain via Wikimedia Commons.

1. This series draws in large parts on the chapters on historiographical perspectives on the digital humanities and network research of my dissertation (Petz 2022). [↩]

---

---