Data imagined as data
Leaving these questions open for now, it will be a central argument here that before we attempt to tentatively answer suggestions about how to better incorporate aesthetic approaches within the sonification of data, we should ask what we understand by data in the first place? In the introduction to a recent anthology, Raw Data Is an Oxymoron, Lisa Gitelman outlines how a broader and more thorough understanding of the vast amounts of data that now make up part our everyday lives is needed; we have to realize that data is always “familiarly ‘collected,’ ‘entered,’ ‘compiled,’ ‘stored,’ ‘processed,’ ‘mined,’ and ‘interpreted’” (Gitelman 2014: 3). Gitelman asserts that the interpretation is often conceived of as a posterior phase of overall data-handling, but that it is nevertheless not obvious how this interpretative activity “haunts its predecessors.” Crucial to our discussion here is her emphasis that “[d]ata need to be imagined as data to exist and function as such, and the imagination of data entails an interpretive base” (Gitelman 2014: 3).
As was the case with the initial example, i.e. the LHC-data serving as point of departure for both exploratory and more ‘communicated’ or musicalized sonifications, that process and venture was explicitly described, in the wording cited above, as derived from “data” and “data sets.” According to Gitelman, we often tend to think of data as “the starting point for what we know, who we are, and how we communicate,” which repeatedly leads to “an unnoticed assumption that data are transparent, that information is self-evident, the fundamental stuff of truth itself" (Gitelman 2014: 2). Gitelman’s discussion and part of the anthology’s overall inquiry is, thus, seeking to highlight and make clearer the ways in which various disciplines’ use of data is based on “how different data sets harbor the interpretive structures of their own imagining” (Gitelman 2014: 3).
This observation is crucial for how we understand the relationship between the sonification of apparently “objective” or “primary” data and the knowledge-generating activity of hearing and listening, i.e. the interpretative act that is usually thought of as “coming after” the initial processing and transcoding procedure. Here we may also turn our attention to the discussion put forth by Veit Erlmann in which he addresses the a priori conditions for listening and meaning, which makes him ask “[w]hy and how certain orders of knowledge [do] make some aspects of our auditory experience more worthy of attention than others” (Erlmann 2010: 18). Erlmann points to how certain conditions must be given “for something to become recognized, labeled, and valorized as audible in the first place” (Erlmann 2010: 18). This sonic contextualization seems to be highly relevant when dealing with the affordances of sonification at large, but even more so when it comes to the transcoding of designated data sets generated by technical apparatuses. According to the above reflections about data as never being neutral or “raw” but always contextualized and negotiated, the non-audible data that one would intend to translate into meaningful information will in most cases, then, be selected from a basic pre-existing idea about the origin of those data, related to the use of certain technologies in specific contexts. Put in another way, we might say that data sonification presupposes an acquaintance with – and judgment of – the data or data sets used in order for the sonification process to be able to bring forth the patterns, significant events, etc. that could eventually lead to meaningful information. Information which could then subsequently enhance and converge (or falsify) rudimentary assumptions, through exploratory use of technology, into purposeful knowledge. This predicament can be said, again, to resonate with the concept of affordance mentioned earlier as something which designates the possibilities for meaningful engagement with and within an encompassing situation. In line with this, Have and Stougaard have recently re-introduced a theory of affordance from a sound studies perspective in their discussion of audio books (Have and Stougaard 2015). Here the authors point out how “affordance comprises a quality that belongs neither to the medium nor to the user, but emerges in interaction between the two” (Have and Stougaard 2015: 88). Furthermore, they argue that the term affordance entails “the possibilities not yet actualized by media users and producers” (Have and Stougaard 2015: 88). Both of these aspects are present when we consider one of the key technical issues at play in sonification processes: the actual coding or transcoding procedures involved.
“Transcoding” is also the term suggested by Sterne and Akiyama as central to describing how “a wide range of possibilities exists for both the manipulation of data and the interface itself” (Sterne and Akiyama: 549). The term is often referred to via Lev Manovich’s discussion of new media technologies and Sterne and Akiyama make use of it here, among other reasons, to put forward some basic considerations about the indexical representation. In a broader sense, transcoding as a concept also is closely related to the above discussion about the contextualization of data. Manovich has advanced the concept in order to argue for a relationship between cultural and computer technological domains, focusing on the possibilities of software programming for handling sound, images, and text in an extremely versatile manner across traditional medial borders (Manovich 2001: 45). Following this, doing more scrutinizing analysis of the level of software programmability used in sonification practices could potentially reveal some of the underlying principles for our interpretative engagement with abstract data sets. However, Adrian McKenzie has argued that we should be skeptical toward the concept of transcoding and claims that it might be an approach that “raise[s] more problems than it solves, since its terms are themselves derived from software design” (McKenzie 2006: 4). By this critique McKenzie means to point out how software design is equally pre-determined in how it renders data according to specific cultural presumptions and institutional circumstances.
Thus, when considering critically the affordances of sonification, one becomes aware of the complex mesh of interrelations that emerges – thinking again about our initial example – between the “objective” data produced through specific technological processes at CERN, the potential information assumed to be contained within those data, and the “official” knowledge generated through interpretative and re-mediating procedures, both individually and collectively by the particle physicists whose personal suppositions of what an electron might sound like can end up shaping the actual expression and aesthetics of the particular sonification of LHC-data (cf. the initial quote above). Here the words of Steven Connor may also resonate strongly in that he suggests how “[s]onification does not sound very far away from personification,” a statement put forth by Connor while explaining his highly skeptical attitude toward a current so-called “conversion hysteria” (Connor 2013: 137). Connor also discusses what he terms “the stubbornly genitive case of sound, its inseparability from the idea of an originating circumstance” (Connor 2013: 151) and identifies a main problem within current notions of sonification, namely a kind of “mysticism of the primal, a set of beliefs that sees translation into sound as a kind of making manifest of the latent truths, of a set of absolute but hidden primal conditions.” “The act of sonification,” he continues, ”is understood as a kind of re-enchantment of the world […] (Connor 2013: 147).
Nonetheless, as highly skeptical of the affordances of sonification Steven Connor appears to be here, we can take this as a cue for moving into the latter part and toward the conclusion of this article. Not least if Connor’s critique is seen in conjunction with how Erlmann has proposed that we revise our understanding of hearing and listening by looking for “a deeper interpenetration of the biological and the cultural” and “complicat[ing] simple tripartite sender-medium-receiver models” (Erlmann 2010: 17). This denouncing of a classical model of communication suggests, among other things, that the sheer physical and material qualities of the media used for transferring data into acoustics should be considered further, even though this analytical task would not immediately seem to be occupied with clarifying the information and knowledge production processes involved in sonification.