Ninth Workshop on the Philosophy of Information

The contrast between the epistemological and computational problems of visualisation provides us with a first line of inquiry for exploring connections between different disciplinary perspectives on visualisation.

For the philosophy of science, the epistemological problem of information visualisation is the central problem. This problem is concerned with what visualisation is and how it is used in the sciences. It deals with the nature of visualisations as a kind of epistemic artefact, and asks how (if at all) such artefacts are related to the reality they are meant to be related to, which epistemic role they play, and whether their use in the sciences is efficient and perhaps even indispensable. The problem of visualisation can be considered, in this perspective, a sub-species of the problem of scientific representation.

In this context, questions include: how visualisations represent (Kulvicki 2010, Bolinska 2016), how they can be used to reason “by proxy” about a system they represent (Suàrez 2004, Contessa 2007), or whether visualisations can generate or convey insights or results that could not, or at least not as easily, be obtained by non-visual means (de Regt 2014, Mößner 2014, Boumans 2016).

The visualisation sciences, understood as the disciplines within computer science that are concerned with all aspects of the computer-aided generation of graphics based on (numerical) data obtained from measurement or simulation (see e.g. Haber 1990), adopt a different —computational rather than representational — perspective that cannot as easily be related to what philosophers of science focus on. When it comes to the topics of insight and understanding, however, we find a more marked thematic convergence: de Regt (2014) questions whether visualisation is necessary for understanding, whereas Chen et al. (2014) challenge the received view that “gaining insight” is the primary purpose of visualisation.

As most questions that are of central importance to the technological practices of information-visualisation and the visualisation-sciences in general remain hard to interpret from a purely epistemological standpoint, one may ask how the computational and epistemological perspectives could be reconciled. We may, however, also wonder whether philosophers shouldn’t instead recenter their attention on the computational problems.

The exploration of potential convergences is valuable for at least two reasons. First, and most obviously, because the increased importance of software in scientific practice, and the distinct value of visualisation for data-intensive sciences, the view that visualisation has become a computational practice can no longer be ignored. Secondly, because one of the main open problems within the visualisation sciences concerns its theoretical foundations (Johnson 2004; Purchase et al. 2008; Chen et al. 2017), there is room for a constructive as well as a critical input from philosophy. A foundational theory of information visualisation that is meant to provide unity and direction to the field, drive progress, and improve the scientific standing of the discipline, needs to engage with epistemological challenges that exceed the engineering challenges of visualisation. Here lies philosophy’s first, constructive, opportunity. At the same time, when a foundational theory is meant as a step towards a more mature science of visualisation, this generates conflicts between the more positivistic tendencies in the visualisation sciences and the more humanistic perspectives on computing and information science. Philosophers can, by building on their prior experience with the question of whether the explanation of knowledge practices requires a unified theory, or are instead better accounted for by situating them within a more diverse, dis-unified landscape, adopt a more critical stance as well.

References

• Bolinska, Agnes. 2016. “Successful Visual Epistemic Representation.” Studies in History and Philosophy of Science Part A 56. Elsevier BV: 153–60. doi:10.1016/j.shpsa.2015.09.005.
• Boumans, Marcel. 2016. “Graph-Based Inductive Reasoning.” Studies in History and Philosophy of Science Part A 59. Elsevier Ltd: 1–10. doi:10.1016/j.shpsa.2016.05.001.
• Chen, Min, Georges Grinstein, Chris R Johnson, Jessie Kennedy, and Melanie Tory. 2017. “Pathways for Theoretical Advances in Visualization.” IEEE Computer Graphics and Applications 37 (4). IEEE: 103–12.
• Contessa, Gabriele. 2007. “Scientific Representation, Interpretation, and Surrogative Reasoning*.” Philosophy of Science 74 (1). JSTOR: 48–68. doi:10.1086/519478.
• Haber, Robert B, and David A Mcnabb. 1990. “Visualization Ldioms : A Conceptual Model Visualization for Scientific Systems.” Visualization in Scientific Computing 74: 93.
• Johnson, Chris R. 2004. “Top Scientific Visualization Research Problems.” IEEE Comput Grap Appl IEEE Computer Graphics and Applications 24 (4): 13–17. doi:10.1109/MCG.2004.20.
• Kulvicki, John. 2010. “Knowing with Images: Medium and Message.” Philosophy of Science 77 (April): 295–313. doi:10.1086/651321.
• Mößner, Nicola. 2014. “Visual Information and Scientific Understanding.” Axiomathes 25 (2). Springer Netherlands: 167–79. doi:10.1007/s10516-014-9246-7.
• Purchase, H C, N Andrienko, T J Jankun-Kelly, and M Ward. 2008. “Theoretical Foundations of Information Visualization.” Information Visualization 4950 (4950): 46–64. doi:10.1007/978-3-540-70956-5_3.
• de Regt, Henk W. 2014. “Visualization as a Tool for Understanding.” Perspectives on Science 22 (3). MIT Press - Journals: 377–96. doi:10.1162/POSC_a_00139.
• Suárez, Mauricio. 2004. “An Inferential Conception of Scientific Representation.” Philosophy of Science 71 (5). The University of Chicago Press on behalf of the Philosophy of Science Association: 767–79. doi:10.1086/421415.