The ‘materiality’ of digital information – what are we talking about? / by Christoffer Andersson

One of Googles data centers in The Dalles, Oregon

One of Googles data centers in The Dalles, Oregon

In this text I will consider what it means to talk about the materiality of digital information. Does it make any sense to talk about the materiality of virtual objects and phenomena? And even if it makes sense, what can it help us understand about digital information?

As one may gather from the colloquial use of the term materiality, it is gesturing towards the physical world which is made up of things, objects and materials somehow. Materiality then, must be about the thingness of things, or the quality of things, the features of things, in some way. What first comes to mind is that we can describe the physical qualities and features of things — a bridge is made of steel girders with this and that tensile strength. While it is certainly the case that the physical and/or biological features of things and objects are important, it is not the mere fact that digital information actually has a physical presence that is interesting to me. Although, it is surely important for lots of reasons that, for example, digital information can be stored as states of transistors in an USB-flash drive instead of boxes upon boxes of punch cards.

Instead the features that is of most interest to me, as a researcher of social phenomena in organizations, is all of those that affect the ability of objects and things to make a difference in the world; how they affect change in organizations and reconfigure our practices. A useful word to use in this context and that capture this ability of affect change is dynamism.

Of course, the specific features that make things and objects dynamic in the world are wildly different depending on what type of thing we are talking about. Rocks in the forest are not very dynamic, not especially loaded with potential to affect the world. Computer programs and robots, on the other hand, may be very dynamic — loaded with potential to change events and processes. What then is it about digital information that produces a different kind of materiality and dynamism than is present (or not present) in rocks?

Paul Dourish and Melissa Mazmanian have written about the various ways in which the theme of the materiality of information has been treated in social research and they identify different types of conceptualizations that has been used. First off, scholars have for a long time investigated the ways in which things and objects carry cultural meaning, symbolic value and do important cultural work. This pertains to the use of digital things as well. Think of all the fuss about what it says about a person if one uses Apple products or not, Linux as an operating system or not, Mozilla Firefox or Internet Explorer. Digital things, like ‘regular’ things, carry cultural importance

Another way is how information systems of various kinds are woven into space; the internet connections in our homes, the cell-phone towers looming tall over desolate forests, and how these infrastructures affect our experiences of those spaces. The availability of certain infrastructures that enable the transmission and propagation of digital information increasingly contextualizes our experience. Think of the annoyance of a shoddy internet connection, or how cellular coverage or not can determine property prices in rural areas.

The third way that Dourish and Mazmanian present is the concern scholars have had with the material conditions of information technology production. There are of course production factors to consider even for digital things. Not just the labour conditions in the factories, or the open wounds on the earth and societal toll that mining of rare earth elements needed for our favorite products create. Also, as Lucia mentioned in her post, the labour that is needed to train algorithms to distinguish a small child from another feature of the road for self-driving cars or to prevent users from viewing graphic violence or pornography on Facebook. Underpinning such algorithms are countless hours of human labour in order to create the data sets needed for so called ‘machine learning’ to be possible in the first place.

The final way they describe is the materiality of information representation. What does an obtuse turn of phrase like that even mean, one might ask? Well, it turns out that the way information is represented is a key aspect of organizational practices for example. The particular form that information takes shape which questions we can ask of it, the kinds of manipulations and analysis it supports and how it can be used to shape our understanding of the world. It matters if information is stored in a relational database or as a scrambled mess of discrete files if we are to extract information in a structured way. It matters if we use visualizations, numbers or text and it matters which format each of these takes.

This last point is what is of most interest to me in my research of automatization in white collar work. It transpires that the materiality of information representation has a lot to do with how flexible organizational processes can be — if algorithms are to be deployed then their materiality matters. That is to say, there are properties and features of algorithms as representations that constrain, enable and shape the way information can be put to use, transmitted, acted on and, importantly, its dynamism. Therefore, specific configurations and forms of algorithms – how they are figured and how they are deployed — really matters.

So, when one talks about the materiality of digital information, one can refer to all of these aspects, and all of these aspects are important to understand if one is interested in how digital technology affect management and organizational practice. Depending on the particular thing or object, some aspects may be more important than others. But materiality is certainly important even when talking about digital information, even though the virtual might seem like the opposite of material.

 Christoffer Andersson, Phd candidate