Techniques and Visualization in the Digital Humanities Need to Address the Needs of Humanists
Humanities approaches can potentially have a large impact on digital scholarship. For example, markup languages, initially considered as helpful tools for the digital humanities, have become the objects of substantial controversy, especially in the late 1990s. For instance, there was the perception that the hierarchical structure of the Extensible Markup Language (XML) was limiting to digital humanists and conflicted with the manner with which humanities scholarship was carried out. However, these discussions did not result in techniques being more grounded in the theoretical principles of the humanities. Indeed, the focus has been how computational, “digital” technologies have affected work in the humanities. Another, perhaps more important question is how the humanities can affect computational techniques, visualizations, and interfaces. That is, can graphical interfaces and digital platforms be developed from and based on methods used in humanities work? Put another way, the focus of humanistic studies need not be exclusively fixated on technological innovations but should shift towards a theory of technology development in design, information modeling, data representations, and interfaces, all embodying humanistic concerns (Drucker, 2012).
For work in the humanities, the objective, mechanistic, positivistic, quantitative nature of standard data processing and visualizations is not sufficient, and in fact precludes humanistic theoretical principles and thought patterns. Even in the domain of visualization, objective graphics, such as street-level satellite imagery or maps from geographical information systems cannot be taken by humanities scholars as simple objective measures. The two areas – quantitative and humanistic – start from different epistemological bases. For instance, texts are generally thought to be objective expressions. However, humanists consider texts to be produced by the act reading itself, and are therefore not deterministic, objective constructs.
As Drucker states, “… this is a critical moment to identify core theoretical issues in the humanities and develop digital platforms that arise from these principles” (Drucker, 2012).
Information visualization, data mining, and other data science techniques have different epistemological foundations than those in the humanities. Graphs and charts are overly objective and do not represent the nuances inherent in the humanities.
A dichotomy also exists between the quantitative, mathematical, and ultimately deterministic methods offered as tools to the digital humanities, and the probabilistic, performative models preferred by practitioners of those disciplines. Advanced models of simulation need to be developed to align more closely with the practices of humanists. Statistical modeling, dealing with probabilities instead of certainties, offers benefits. However, the goal is to model ambiguity within humanities investigations, and probability theory, integral to statistical models, can only imperfectly approximate such ambiguity. In addition, the corresponding visualizations still reflect deterministic outcomes. Consequently, great opportunities exist to develop visualizations that reflect ambiguous and partial knowledge.
Another disconnect between mathematical and computational tools and the needs of the digital humanities arises in the context of space and time. For humanists, the most standard spatial visual representation, maps, which are usually considered to be objective structures, are in fact human constructs. Traditionally, the nuances of social, cultural, and phenomenological expressions are not properly represented in standard spatial and/or temporal visualizations. Typical “objective” maps, such as those used in GIS systems or obtained from satellite imagery, obscure linguistic boundaries and other social groupings. As another example of this inadequacy, a straight line – or even a great circle – connecting two geographic locations on a map, may be used to indicate a route for sending a letter from one location to the other. However, even this straightforward representation is problematic because it implies a direct, smooth movement from one point to another, and does not take into account the reality of detours, checks, delays, travel at different speeds, etc. which are normally part of every mail delivery. Similarly, to use Drucker’s illustrative example, a fully interactive and high-fidelity virtual 3D model of inscriptions in the Roman forum do not capture the “poetics of space” and the social relations that are part of the experience of this forum. The high-fidelity interactive model may be much more compelling than the fragmentary and incomplete “humanistic” model, fraught with complexity. However, the latter model deals with exactly those issues faced by humanists.
The same difficulties arise in temporality, or notions of space, as time, like space, is greatly affected by cultural circumstances. Life expectancy, time measuring devices (such as clocks) and artificial lighting greatly affect people’s perception of time. Treating time as one of the four physical dimensions (in addition to the three spatial dimensions) does not take these nuances into account and is therefore inadequate from the perspective of the humanist. Therefore, like space, the experience of temporality, in all its multidimensional facets, requires a new graphical language for analysis and dissemination.
To summarize, lived experience, the subject matter of the humanities, is distorted when mapped onto the linear scales required by standard quantitative analyses and visualizations. For the humanist, knowledge is interpretation, and, consequently, standard techniques and visualizations, which are often considered as self-evident, do not completely suffice in humanities work. As Drucker states, “[h]umanistic theory provides ways of thinking differently, otherwise, specific to the problems and precepts of interpretative knowing – partial, situated, enunciative, subjective, and performative” (Drucker, 2012). Therefore, for computational techniques to adequately address the needs of humanists, and to be properly called “digital humanities tools”, those principles employed by humanists in their work must contribute to the development of new methods that meet those needs. For software developers, computer scientists, mathematical modelers, and practitioners in other technological disciplines, these are difficult challenges, but provide important and exciting opportunities. As Drucker puts it at the end of her essay, “[t]he question is not, Does digital humanities need theory? but rather, How will digital scholarship be humanistic without it?” (Drucker, 2012).