Towards Feminist Data Production; A Case Study From Comics

Introduction

This paper will take my ongoing doctoral research designing resources for analysing individual graphic novels as a case study and starting point to discuss how we can produce data in the Humanities which can be called ‘feminist’. It will engage with the significant debates on this topic which are emerging in the Digital Humanities (Clement, 2016) (Drucker, 2012) (Losh et al., 2016) (Posner, 2016) (Rhody, 2016), particularly the panel discussion on creating feminist infrastructure at DH 2016 (Brown et al., 2016). Marking up images and designing ways in which to make them communicate with a comic’s words are highly subjective enterprises. By explaining how I dealt with this issue in my research, this paper will outline how the subjectivity involved in creating our own data structures and ontologies, and these elements’ inherent statuses as arguments about data, is a strength, an affordance, of such an approach, as well as something that embraces the situatedness and plurality of data that feminists in the sciences (Haraway, 1988) (Irigaray, 1985) have advocated for and which, more recently, have been advocated for in the Digital Humanities. Although my case study focuses on contemporary comics, this paper will explain how some of the resulting principles can be employed by data creators today in other disciplines and the GLAM sector.

Background and method

As a Digital Humanist operating outside of a specific department or centre, the question that I hear most often about my work is how I make the resources which I create objective, how I avoid my datasets merely reflecting one individual’s interpretation of a text. But when I mark up an image in a certain way, or use a database structure to reflect the rhizomatic structure of a graphic novel, I do so not because it is ‘appropriate’ or a ‘good fit’, but because that is an argument I want to make about comics and their meaning mechanism, and by applying that algorithm to the dataset that is the comic, I articulate that argument, mobilising it and making it available for evaluation. I am not trying to enact as little violence as possible to a text; I am making an argument about it. This idea that datasets, data structures and algorithms are arguments that are made about texts or other objects of study is relatively well-established in Digital Humanities (Ramsay and Rockwell, 2012) but it is more often framed as a caution than an opportunity.

Working on contemporary comics, where there is no pre-existing database, and no automated or straightforward tagging of images, it would be easy to see mark-up or tagging as a hurdle, and a problematic one at that, given the fraught nature of remediating pictorial information into values that can be entered into a database. But, although I must design my own data tags and ontology, I do not have a mandate to preserve, gatekeep, or distribute otherwise inaccessible data since my objects of study are widely available. Focussing on individual texts, too, affords me time to spend designing tailored data tags and ontologies. I do not need to preserve, that is my freedom; I cannot be ‘objective’, that is my strength. My objects of study are relatively small; that means I do not have to be singular, I can be multiple.

Digital approaches, especially to comics (Walsh, 2012) (Dunst et al., 2016), often rely on a single categorisation of each entity or attribute. This paper will argue, rather, that our databases ought to be multiple. Rather than text-mining, a metaphor which suggests the removal of gratuitous material, I would encourage thinking of this practice as data curation, or rather, curations. Consider the analogy of a virtual museum with access to a complete catalogue of material - for is this not like our complete texts? -where anybody can hang the material in whatever way, in whatever ways, they choose. Different paths through the information can be curated, different logics created, retaining the plurality of signification that each piece holds, resisting positivism. This may well tell us as much about our hypothetical hangers as about the hung objects, but therein lies an opportunity. If curation - like datasets, data structures and algorithms - is argument, then why not bring multiple perspectives into conversation? And if we can represent data multiplicitously, we can do it investigatively. By creating multiple ontologies and data tags it is possible to embrace Brian Massumi’s judgement, “[t]he question is not: is it true? But: does it work? What new thoughts does it make possible to think?” (Deleuze and Guattari, 1987: xv)

Since datasets are arguments, marking up the same data differently allows researchers to avoid asserting single values for complex, or simply ambiguous, pieces of information. It also liberates us to encode arguments we disagree with, or at least that we concede are problematic, in order to better understand or critique them. The reduction of gender to a binary, for instance, has been highlighted as an issue in quantitative approaches (Clement, 2016). This, of course, drags up familiar tensions between anti-essentialism and feminism, but if we can encode different modes of representing gender - or any other ‘attribute’ which is better represented on a spectrum -including the reductive binary mode, we can maintain the plurality of our data, whilst retaining the possibility to see how the text subverts such a binary categorisation; we bring the text to bear on the theory and in so doing, better understand the theoretical position of the text. Image tagging can operate in a similar way, by tagging the same pictorial signifier in multiple ways we can tag with an intention to investigate, not merely minimise violence to the text. By contrasting and combining different ontologies it is possible to shed light on our texts and to allow our texts to shed light on our ontologies, all the while fracturing any notion of computational methods as objective black boxes by foregrounding their artificiality.

Conclusion

Embracing a conceptualisation of Humanities data as complex and plural, this paper will use examples from my own research remediating graphic novels into databases to demonstrate how deploying multiple tags and multiple ontologies not only instantiates a more feminist approach to data but is actually a productive methodology for analysing texts. It champions not the analysis of datasets, but rather an analysis by datasets. As Laura Mandell said in Krakow, we need “metadata built for thinking, not sustainability.” (Brown et al., 2016)

Bibliography

Brown S., Clement, T., Mandell, L., Verhoeven, D., Werni-mont, J. (2016). “Creating Feminist Infrastructure in the Digital Humanities.” Digital Humanities 2016: Conference Abstracts. Jagiellonian University & Pedagogical University, Krakow, pp. 47 - 50.

Clement, T. (2016). “Where is Methodology in Digital Humanities?.” In Gold, M. and Klein, L. (eds), Debates in the Digital Humanities. University of Minnesota Press, pp.

153 - 75.

Deleuze, G. and Guattari, F. (1987). A Thousand Plateaus. trans. by Brian Massumi. London: Continuum.

Drucker, J. (2012). “Humanistic Theory and Digital Scholarship.” In Gold, M. (ed), Debates in the Digital Humanities. University of Minnesota Press, pp. 85 - 95.

Dunst A., Hartel, R., Hohenstein, S., Laubrock, J. (2016). “Corpus Analyses of Multimodal Narrative: The Example of Graphic Novels.” Digital Humanities 2016: Conference Abstracts. Jagiellonian University & Pedagogical University, Krakow, pp. 178 - 180.

Haraway, D. (1988). “The Science Question in Feminism and the Privilege of Partial Perspective.” Feminist Studies, 14(3): 575 - 99.

Irigaray, L. (1985). “Is the Subject of Science Sexed?” trans. by Oberle, E. Cultural Critique, 1: 73 - 88.

Losh, E., Wernimont, J., Wexler, L. and Wu, H. (2016). “Putting the Human Back into the Digital Humanities: Feminism, Generosity, and Mess.” In Gold, M. and Klein, L. (eds), Debates in the Digital Humanities. University of

Minnesota Press, pp. 92 - 103.

Posner, M. (2015). “The Radical Potential of the Digital Humanities: The Most Challenging Computing Problem is the Interrogation of Power.” Impact of Social Sciences Blog. Web. 15 Sept. 2016.

Ramsay, S. and Rockwell, G. (2012). “Developing Things: Notes toward an Epistemology of Building in the Digital Humanities.” In Gold, M. (ed), Debates in the Digital Humanities. University of Minnesota Press, pp. 75 - 84.

Rhody, L. (2016). “Why I Dig: Feminist Approaches to Text Analysis.” In Gold, M. and Klein, L. (eds), Debates in the Digital Humanities. University of Minnesota Press, pp. 536 - 39.

Walsh, J. (2012). “Comic Book Markup Language: An Introduction and Rationale.” Digital Humanities Quarterly, 6(1).