For this reason, in this webtext, we make purposeful, interdisciplinary connections between visualization (broadly conceived) and translation, showcasing how rhetoric and composition, technical communication, and digital humanities connect language fluidity and visualization for a variety of purposes, including intercultural communication, multilingual and multimodal composing, and computational visualization. As a result, scholarship across fields has focused on ethical representations within visual products and the ethics of process and production of those visualizations.
-
Technical Communication
Within technical communication, visualization and translation have been connected to issues of data representation, ethics, and intercultural communication (among others).
- Although some scholars in technical communication ask questions about methodologies related to big datasets (Graham et al., 2015), most discussions of data visualization are currently based on information design. Even before the proliferation of complex and interactive data visualizations, scholars in technical communication studied information design, or the ways words and visuals come together to increase audience comprehension (Schriver, 2003). As Paul Kußmaul (2005) explained, visual representations like "Pictograms are sometimes used on their own to refer to things, states of affairs, emotions etc. or even for symbolizing speech acts" (p. 380). When technical communicators encounter complex information that cannot be presented graphically, they often utilize their own design skills and processes to visually represent information (Sánchez, 2017). Because technical communicators simultaneously accommodate complex information and frequently design visualizations of that information, scholars have increasingly sought to answer questions around the ethics of representation (Buehl, 2014; Ross, 2017). Sam Dragga and Dan Voss (2001) reviewed the "inhumanity of technical illustration" by examining bar and line graphs, drawn diagrams, and pie graphs that represented death from events like logging accidents, mattress fires, and baby walkers. Through this analysis, Dragga and Voss argued that the significance of these deaths appear to be "ordinary … facts like all other facts" (p. 269) through their visual representation. By simplifying human deaths into data points, the visualizations became inhumane cruel pies. As visualizations become exceedingly intricate and complex, Dragga and Voss's call to "promote a humanized and humanizing understanding of technical subjects" (p. 272) remains applicable.
- Building on the ethics of representation, technical communication scholars have considered the persuasive principles of visualizations, particularly in relation to intercultural communication and social justice (Jones, 2015; Kostelnick, 2016). For intercultural and multilingual technical communication researchers, visualization is both a process and a product. Successful translation and localization are directly linked to visualization and encompass more than visual design. The process of translation and localization itself has been described as a visual, embodied activity that requires translators to engage in the transformation of language beyond alphabetic genres.
- Scholars working in intercultural technical communication, and technical translation specifically, have historically turned to visualization as a way to present and adapt information across languages. For example, through her study of technical translation students' translation diaries, Anne Ketola (2015) demonstrated that translators frequently "negotiated the relationship between modes" (p. 33), including specifically the interactions between alphabetic and visual information, as they translate. As Ketola explained, "translating today often involves engaging with multimodal material," including "written language, spoken language, images, etc." (p. 13). This argument is echoed by research on localization, which emphasizes the fact that translations, particularly in contemporary contexts, cannot be fully localized to meet the expectations and needs of specific communities without considering visual design (Rose et al., 2017; Sun, 2012). In their study intended to improve the design and usability of a document designed to help immigrant patients sign up for health insurance, for example, Emma Rose et al. (2017) explained that the visual complexity of a health-insurance guidebook negatively impacted access to health-related information for multilingual immigrant communities. Similarly, in their study of medical interpreters' translation strategies, Laura Gonzales and Rachel Bloom-Pojar (2018) pointed to the ways that multilingual communicators sometimes rely on sketching, drawing, or creating on-the-fly visuals to facilitate communication between English-speaking healthcare providers and their multilingual patients. In their work on using visualizations to navigate "translation spaces," Bloom-Pojar and Danielle DeVasto (2019) concluded that "visuals are often deployed as tools to simplify communication with publics and patients who communicate differently from health professionals and public health officials. When visuals are used to 'transcend' language, or avoid language differences, a key component is disregarded: Visuals are not necessarily transparent, objective, or universally understood." Finally, in their study of "untranslatable" terminology, Gonzales and Rebecca Zantjer (2015) identified visualization as a common strategy employed by multilingual communicators to translate difficult terminology for a variety of audiences.
-
Rhetoric and Composition
Rhetoric and composition researchers connect translation and data visualization as multimodal modes of interpretation.
- Like translation, other multimodal production work has been shown to involve more than just technical skill, instead requiring "careful attention to both traditional and technological rhetorical considerations" (Sheppard, 2009). Such rhetorical and technical considerations in multimodal writing happen when data is visualized, even though all multimodal writing is not data visualization. Data visualizations specifically and multimodal writings in general are not just end products, but are also generative processes mediated by identities and physical bodies (Arola & Wysocki, 2012). This is especially true when multimodal composition is enacted in multilingual contexts, when writers have to navigate linguistic and cultural elements while also engaging in cross-cultural design (Fraiberg, 2010; Horner, Selfe, & Lockridge, 2015).
- Considering the typical uses of data visualizations that "represent the world in numbers and then represent numbers in graphic form," scholars like Seth Long (2016) are asking: How can a rhetorician approach data visualization? Both Long and Derek Mueller (2017) argued for data visualization as inquiry generation. Mueller's Network Sense: Methods for Visualizing a Discipline articulated visualization as a methodology for rhetoric and composition that unites numerical data mined from "pattern-amplifying devices" like journal articles, conference proceedings, syllabi, and more with digital humanities methods like distant reading and thin description to create "incomplete but nevertheless vital glimpses of an interconnected disciplinary domain focused on relationships that define and cohere widespread scholarly activity" (p.xii). These vital glimpses come in the form of visualized patterns—like word clouds, citation frequency graphs, and maps of scholarly activity—generated by databases and computational software (like Tableau) and platforms (like Google Charts). As Mueller argued:
The visual models are not proofs, finally, but provocations; not closures, but openings; not conclusions or satisfying reductions, but clearings for rethinking disciplinary formations—they stand as invitations to invention, to wonder, as catalysts for what Ulmer described as "theoretical curiosity" (p. xii). (p. 4)
- Instead of only using hyperlocalized accounts like ethnographies and case studies to get a baseline of "complex, distributed disciplinary activity" (Mueller, 2017, p. 8), visualizations take complexity and "[hold] the text at bay so that we might see it instead as a semantic network with concentrations of terms coalescing throughout it" (p. 5). Thus, visualizations provide "thinned out" (p. 6) or simpler patterns that compel researchers to ask, "What next?"
-
Digital Humanities
Digital humanities research offers a systematic method for structuring data to be visualized and critiques the limitations of those structuring methods.
- Data visualizations often require researchers to "clean," or manipulate, data in a way that platforms and software can read and work with (Rawson & Muñoz, 2016). In initial phases of computational data collection processes, researchers categorize, collapse, and modify subjects in order to make information systems process data more efficiently. Sometimes cleaning can be as straightforward as making a date format (mm/dd/yyyy and m/d/yy will not be recognized together) consistent across a database. Because these visualizations require front-loaded data manipulation and the use of digital tools and software, these kinds of visualizations occur at the end of research, often as a research finding, repository, or archive. Simplifying data by cleaning, like any kind of data selection, is also social and political. Katie Rawson and Trevpr Muñoz argued that the cleaning paradigm in data visualization assumes an underlying, correct order. Although cleaned data appears tidy grouped into rows or columns or neatly-delimited records, this tidiness "privileges the structure of a container rather than the data inside it" (Rawson & Muñoz, 2016). Thus, as digital humanities scholars argue, databases frequently privilege computer systems more than data by requiring specific logics.
- By using computational procedures from data science to manipulate and order information, lived experiences may be oversimplified or abstracted to meet technical specifications of software. Furthermore, the visual outputs of these computations frequently appear as neutral presentations of coded information, while their reality is a human-authored rhetorical construction with immense persuasive power (D'ignazio & Klein, 2020). But who authors these technical specifications and what logics are privileged in that coding? Catherine D'Ignazio and Lauren F. Klein (2020) argued that many standard practices in data science, like binary classification systems and cleaning, "become naturalized as 'the way things are.' This means we don't question how our classification systems are constructed, what values or judgments might be encoded into them, or why they were thought up in the first place" (p. 104). The reduction and simplification that makes data visualizations possible can also be harmful, reinscribing colonial frameworks that have historically been used to undermine and oppress cultural, community-driven practices (Agboka, 2014; Haas, 2012; Patel, 2015).
- Data visualization relies on the interpretations of researchers and their computational instruments. Maggie Walter and Chris Andersen (2013) reiterated that "quantitative data play a powerful role in constituting reality through their underpinning methodologies by virtue of the social, cultural, and racial terrain in which they are conceived, collected, analyzed, and interpreted" (p. 9). In other words, when these subjective positions, or what Natasha Jones, Kristen Moore, and Rebecca Walton (2016) called positionality, are unacknowledged in research design, such methods have the potential to reinscribe systems of oppression. Practicing data visualization as a set of integrated strategies can increase opportunities for collaboration, encourage ethical collaboration, and lead to ethical attribution and representation.
- This is not to say that computational methods do not have a place in culturally sustaining research. Instead, to design and conduct research that accounts for positionality and humanizes subjects, critical digital humanities scholars like N. Katherine Hayles (2007) argue that there are opportunities to expand database structures: "the great strength of database, of course, is the ability to order vast data arrays and make them available for different kinds of queries" (p. 1604). If data does need to be cleaned, critical DH suggests collaborative methods like open database design which make those categorizing and collapsing processes transparent and also editable. By iteratively designing and redesigning a database collaboratively, structures give way to relations or what Mueller (2017) called "media objects [that] pluralize monolithic logics for expression" (p. 51).