Possible Futures of The Digital Humanities
It is expected that digital technology, hardware and computational architectures, storage capabilities, infrastructure, and algorithms will make significant advances in the foreseeable future. As discussed below, innovations in leading-edge visualization, human-computer interaction, and particularly in machine learning algorithms, will continue to contribute to the transformation of the digital humanities.
Digital humanists must therefore remain current on new trends and developments in machine learning and quantitative methods. As an example, although topic modeling is ubiquitous for understanding the structure of text corpora, newer methods may offer substantial benefits. Spectral clustering is one such alternative. Spectral clustering is an unsupervised data categorization method that determines the complete hierarchical structure of a corpus, which can then be visualized. In this way, this new method is conducive for studying multiple corpora. Spectral clustering is some ways simpler than topic modeling. Whereas the latter uses LDA (latent Dirichlet allocation) and requires parameters to be adjusted, or tuned, spectral clustering does not require such tuning. Additionally, the number of latent topics does not need to be pre-specified. Spectral clustering is also deterministic in that the algorithm always produces the same result (Arnold & Tilton, 2019). As a further example, the structure of texts may be visualized with dimension reduction techniques. With this approach, documents or topics do not need to be forced into discrete categories (Arnold & Tilton, 2019).
However, these computational advances are not the main drivers of progress in the digital humanities. The field will continue to grow and develop in concert with developments in the (non-digital) humanities. Existing technologies can be applied to the humanities in new ways so that the latter more fully leverages the tools and techniques that have already been developed, tested, accepted, and adopted in a variety of other disciplines. Alternatively, Mark Hedges of King’s College London expresses it as follows: “The potential of digital research methods in the humanities and cultural heritage sectors is reliant not on the emergence of new technologies or discoveries, but rather on the application of existing technologies” (Hedges et al., 2019). But humanities questions themselves are seen as the main source of innovations in the digital humanities.
