Sheer Hardware: Material Computing in the Work of Martin Howse and Ralf Baecker
The emergence of ‘creative code’ in the media arts and design has run parallel with a creative investigation of the materiality of digital media. Developing the author’s previous work on transmateriality, this paper considers material computing in the work of Martin Howse and Ralf Baecker. These works emphasise the mineral substrates of the digital computer, and seem to realise Friedrich Kittler’s proposal in ‘There is No Software’ (1995) for “sheer hardware” - a construct that taps the immanent connectivity of the material world. After tearing down to primitive electro-mineral components, Howse and Baecker reconstruct computing here in a form that is both unfamiliar and timely; it not only affirms the materiality of computing machines, but shows that material computing entails a model of ‘distributive’ agency more complex (and contingent) than programmer and programmed.
For media arts practice, code is now both a ubiquitous tool and a cultural marker. ‘Creative code’ celebrates the potential of a hands-on approach to software (Maeda 2004), while methodologies such as critical engineering focus on the political leverage that code affords (Oliver, Savičić & Vasiliev 2011). Yet while code has been on the ascendancy in digital art and design, we have seen within the same field a persistent preoccupation with the ‘hardware’ to code’s software: the materiality of computing and media technology. In previous work I have outlined some of the features of this practice, and begun to theorise the paradoxical materiality of digital media that it reveals. ‘Transmateriality’ is an attempt to theorise digital media as everywhere and always material, despite often behaving otherwise (see Whitelaw 2011, 2012). As Matthew Kirschenbaum writes: “computers ... are material machines dedicated to propagating a behavioral illusion, … a working model, of immateriality” (2005). Characteristic strategies of ‘transmaterial’ practice include transduction - exposing the translation or mapping of patterns and data between substrates - and a turn towards specificity, and away from generalising infrastructures such as the digital screen (Whitelaw 2011). Here I extend this work to address the materiality of computation itself, through a study of the work of two contemporary media artists, Martin Howse and Ralf Baecker.
The material turn in the media arts runs in parallel with the ‘new materialisms’ emerging in cultural theory and philosophy, though, as Jussi Parikka argues, there is some productive entangling to do in linking media theory with materialist philosophy. New materialism, he writes, “is not only about intensities of bodies and their capacities...” but is in fact “already present in the way technical media transmits and processes ‘culture’” (Parikka 2012: 95). This entails looking beyond things to more ephemeral phenomena - what Parikka calls the “real but weird materialities” of media: “not only touchable objects, but also modulations of electrical, magnetic, and light energies, in which also power is nowadays embedded”. Computation is exactly such an ephemeral phenomenon, and certainly one where power is firmly embedded. Lev Manovich characterises a present where “software takes command” (2013). Meanwhile Marc Andreessen has gleefully declared that “software is eating the world”, celebrating the economic “disruption” wrought by the increasing reach of computation (2010). To suspend the (necessary) political responses momentarily and dwell on Andreessen’s figure of speech, we might ask how this happens - how ‘software’ (not ‘hardware’) eats the world? To eat the world is to be in it, surely; and not surprisingly we see an increasing attention to the material aspects of computing and communication - to sprawling, energy-hungry data centres, and the webs of telecommunication networks (see, for example, Blum 2012), as well as the toxic residues of e-waste. These traces are the scats and footprints of computing - they put us on the trail, but don’t get us much closer to the ephemeral creature itself.
Artistic practice offers both practical specificity and speculative reach, and in the experiments of Martin Howse and Ralf Baecker, we begin to approach the weird materiality of computation itself. Their work offers a focused examination of how software is in the world; they pursue something like media archaeology, excavating the material and technological conditions of computing, and in a way re-staging its emergence. What emerges is quite unlike computing as we know it in an everyday sense; and more, it undermines the conceptual model of the Turing machine that underpins traditional digital computation. Beyond simply ‘grounding’ computing in matter - a sort of necessary substrate - we get a sense here of a computing that is richly entangled and embedded in the world.
There is No Software
In declaring that ‘There is No Software’, Friedrich Kittler marks a useful starting point for this investigation (Kittler 1995). Most readings of Kittler’s essay focus on the argument in its first half, and the core proposal of its title. Kittler argues that “written texts ... do not exist anymore in perceivable time and space but in a computer memory’s transistor cells”. In a “descent” through the architectures and abstractions the computer, we come to rest at hardware. Here, not only texts but software - executable code - also disappears: “All code operations, despite their metaphoric faculties such as ‘call’ or ‘return’, come down to absolutely local string manipulations and that is, I am afraid, to signifiers of voltage differences”. Kittler goes on to critique the layers of abstraction that obscure this base reality, from operating system to graphic user interface, which “hide a whole machine from its users.”
Kittler goes on to consider the implications and limits of the material basis of computation. If computation is fundamentally a matter of physical devices - if representational capacity is constrained, for example, by the word-length of the CPU - then what physical limits do the software abstractions of modern computing conceal? What attributes of matter make it a suitable substrate for computation? Kittler uses Michael Conrad’s theorisation of “programmability” (Conrad 1995). In Conrad’s formulation “structurally programmable” systems have stable, discrete, independent elements, and the connectivity between these elements is necessarily limited. Kittler sees a disjunction here, between the “avalanche of real numbers” presented by the chaos of the world, and the upper limits of the programmable. To “keep up with” the complexity of the real, Kittler seizes on the information-processing power of Conrad’s “nonprogrammable” systems:
Precisely this maximal connectivity defines nonprogrammable systems, on the physical side, be they waves or beings. That is why these systems show polynomial growth rates in complexity and, consequently, why only computations done on nonprogrammable machines could keep up with them. In all evidence, this hypothetical but all too necessary type of machine would constitute sheer hardware, a physical device working amidst physical devices and subjected to the same bounded resources. (Kittler 1995)
After descending through the abstractions of software to the “ground” of the material,
Kittler’s essay ends with this vague but provocative notion of “sheer hardware”; a sort of material or “no-software” computing. Yet as nonprogrammable systems, it seems such systems would be unlike computers as we know them. Kittler’s speculation trails off here; to continue this investigation we can look to the practices of Baecker and Howse, whose work mirrors Kittler’s descent from software to hardware. Through speculative experiments and practical implementations, these artists provide test cases in material computing. As well as fleshing out Kittler’s sheer hardware, they demonstrate some of its implications, as the computer is torn down and rebuilt literally from the ground up. Both Baecker and Howse adopt what I will term proto-digital techniques. These are often precursors of digital components that arose in the analog electronics of early twentieth century, before eventually being miniaturised into the modern integrated circuit that Kittler describes. ‘Proto’ here refers simply to a preceding technology; but I also intend it with a sense of the unknown potential of the prototype, the not-yet-fully-formed. Winding back computing to something more basic enables the artists to rebuild it in a strange and timely new form.
Ralf Baecker - Irrational Computing
Irrational Computing / Documentation from Ralf Baecker on Vimeo.
Fig. 1: Ralf Baecker (2011) ‘Irrational Computing / Documentation’.
Ralf Baecker’s Irrational Computing looks something like a scientist’s workbench: five ‘modules’ - transparent constructions of electronics and crystals, glimmering, clicking and buzzing (Baecker 2011) (Fig. 1). Nothing here resembles a computer; no screen, no keyboard, not even a silicon chip or hard drive mechanism. Tiny lights flash, and a chaotic texture of electronic clicks and chirps fills the room; nothing here is legible, no text, no image. What kind of ‘computing’ is this? Baecker explains the work as a process of zooming in, parallel to Kittler’s descent to sheer hardware. As Kittler observes, the miniaturisation of the integrated circuit renders the site of computation imperceptible; here Baecker reverses this process, performing “an extreme zooming-in on the smallest ‘physical’ units of digital processes”. In particular, Irrational Computing focuses on the mineral substrate of digital computing, crystals such as quartz, silicon, and silicon carbide. The electro-physical properties of these materials make the modern integrated circuit possible. Here the crystals too are zoomed up; instead of nano-scale wafers hidden deep in black boxes, we see actual rocks: a glowing tray of clear quartz, chunks of grey galena, and at the centre of the installation a dark lump of silicon carbide encircled by probing arms. An array of 64 iron electrodes applies pulses of current to this crystal, triggering tiny flashes and clicks, like sparks of lightning within a thunderhead.
Each of the five modules in the work is an independent electro-mineral unit; these are interlinked to form the chaotic ensemble of the work as a whole, which Baecker terms a “primitive macroscopic signal processor”. Each module is a study in a specific mechanism; through material experimentation and the reconstruction of archaic technologies, Baecker investigates Parikka’s “weird materialities” (2012: 97) of the digital. Module I is the silicon carbide ‘display’, described above. The contrast with our familiar glowing rectangles could not be more stark, especially as computer screen densities increase and individual pixels become imperceptible. Here the display is a brute, electro-physical mechanism, and the pixel is not a formal abstraction but a material event. And yet the structural function of the “display” is maintained: signals from the other modules are routed here to emerge as audio-visual patterns.
Module II is the Coincidence Detector: two parallel glass tubes sit atop exposed electronics; four tiny LEDs blink and chirp. Here, Baecker recreates a notable scientific instrument, invented in 1930 by Bruno Rossi (see Bonolis 2011). Walther Bothe and Werner Kolhörster had, in 1929, detected cosmic rays using two parallel Geiger-Muller tubes; these tubes had been devised in order to detect ionising radiation. By isolating coincidences between the two tubes - events where both register particles simultaneously - Bothe and Kolhörster were able to detect the impact of cosmic rays. In 1930, Rossi improved on this design by using an electronic circuit to detect coincidences, dramatically improving the resolution of this “cosmic ray telescope”. At the same time Rossi invented the first electronic AND circuit - a key element of binary digital logic, and one of the four types of Boolean logic gates now packed into integrated circuits in their millions. The proto-digital emerges here as a sensor, a device for sorting and amplifying tiny physical events into discrete binary states of true and false. For Baecker, the coincidence detector is not so much scientific instrument as signal generator, delivering an “absolutely non-deterministic data stream” (Baecker 2012). Functional, digital computing goes to great lengths to insulate itself from its physical environment, to maintain the integrity of its internal logical states; by contrast, Baecker here cracks his “irrational computer” wide open, extracting binary data from the surrounding cosmic flux.
Similarly Module III, the Crystal Field Oscillator, adopts key proto-digital techniques but reroutes them towards a chaotic, immanent materiality. In this module, crystals of Rochelle salt (potassium sodium tartrate) are clamped onto a flat surface, and strung over with copper wires. Rochelle salt is piezoelectric - that is, it generates electrical current when subjected to pressure. Like other piezoelectric materials, the reverse is also true; the crystal physically deforms when subjected to a current. These properties give rise to crystal oscillators, components that provide a stable frequency critical to technologies including radio. They now play a central role in digital computing, as they provide a “clock” signal, slicing time into discrete steps for the organised execution of digital logic. Rochelle salt was used in the first ever crystal oscillator, invented by Alexander M. Nicholson at Bell Telephone Laboratories in 1917. Here Baecker elaborates the elementary crystal oscillator into an unruly macroscopic field: “the crystals are stabilised / amplified by a resonator circuit with an inverted Schmitt trigger. It is the same circuit you need to drive the quartz crystal of a microcontroller or CPU. This circuit kind of jumps into the resonant frequency of the connected crystal, but it is not very stable. The frequency sometimes jumps out...” (Baecker, correspondence 2012). Baecker wrings a sort of unstable stability from this field of crystals, as they push and squeeze each other, inducing and responding to the surrounding current. This signal maintains its function however, providing an aberrant ‘clock’ for the display of Module I.
The Shot Noise Generator of Module IV is another chaotic signal generator built from proto-digital electronics. Here the units are ‘Cat’s whisker’ detectors - each a Galena crystal with a fine, springy electrode touching its surface. This component, used in early (crystal) radios, is a point-contact diode: the junction between wire and crystal conducts electricity in only one direction. In fact, this device is the first ever semiconductor, dating to the early 1900s. Baecker uses six of these crystal diodes, in two roles: three act as noise generators, according to the principle of ‘shot noise’. This phenomenon relates to the discrete nature of electric charge; in very large numbers, electrons can be abstracted into a single homogeneous flow made up of billions of randomly fluctuating particles. At very low levels of current, where only a few particles pass, this randomness becomes relatively significant. Like the Coincidence Detector, the galena diodes here act as noise sources, amplifying the random flow of electrons into binary fluctuations. This noise is then ‘computed’ by three more crystal diodes acting as an AND gate, switching on only when all three noise sources are in agreement.
In electronics, a phase locked loop (PLL) is a timing control system - a device that generates or regulates a regular temporal pulse. In digital integrated circuits, for example, PLLs are used to maintain the synchronisation of distributed clock signals with reference to the master clock. Baecker’s Phase-Locked Loop is a set of oscillating quartz crystals, each connected to the next in a loop. Each crystal module includes a logical XOR gate that combines its own output with that from the previous module: only when these signals are out of sync does the module pass voltage on to the next crystal in the loop. The result is “a very random signal flow” and as Baecker admits “probably not a real PLL” (correspondence 2012); as in the crystal field oscillator, Baecker adapts a protean digital element but coaxes it into generative noise, rather than regulated stability.
As shown in Fig. 2, the five modules are linked, with the display of Module I in a central role. The display takes in two signals: the crystal oscillator of Module III controls the ‘writing head’ or active electrode of the display; the other carries the source to be transmitted - either the shot noise generator, the phase locked loop, or the output of the display module itself. These three inputs are switched by the output of the coincidence detector (Module II). The unstable outputs of all the modules are folded, switched and intermodulated - but with the crisp edges of discrete logic, rather than analog blur. For each module here generates what Baecker calls “data” - digital streams of voltage fluctuations - from material flux. As such these discrete signals can function as logical instructions - for example routing signal within the display, or switching sources. In other words, the signals are both hardware - electrical charges in wires - and software, or code. The Boolean logic of digital computing emerges here, unsteadily, wobbling like something newly born.
Martin Howse - Earthcodes
Martin Howse’s Earthcodes project is an expansive project, both practical and speculative. In the words of the artist, it “proposes an intentionally literal, artistic series of experimental situations which explore the notion of an earth computer, a computational device inscribed or doped … on the earth substrate itself” (Howse 2013a). Howse’s practice as a whole spans experimental audio, electronics, hardware and performance; his preoccupations include psychogeography and “psychogeophysics”, as well as “asking the question of where precisely the plague known as software executes” (Howse 2013b). In the Earthcodes projects, Howse investigates this “site of execution”: “where is the precise point of transition between ... raw material substrate and some kind of inscribed logic which can be bootstrapped into complex symbolic manipulation” (correspondence 2012). Like Baecker, Howse begins by reasserting the mineral substrate of computing: “substrate interfaces with code, yet this set of symbolic, linguistic and logical operations denies the being-substrate, just as the carrier of any signal is erased by the receiver” (2013a). As Howse departs from this point via the conceptual device of the “earth computer”, his work investigates the conceptual and practical limits of digital materiality and its relation to the wider world.
Howse’s Earthboot is a startling practical experiment in the creation of an earth computer. The premise of the work is straightforward: “earthboot enables almost any computer to boot straight from the earth, sidestepping dirty mining actions, and the expensive refining and doping of raw minerals; thus avoiding environmentally wasteful production techniques for the construction of data bearing devices”. The earthboot device is a small trowel-like circuit-board, with two copper conducting plates at its pointed end, and a USB cable emerging at the other (Fig. 3). Video documentation shows Howse crouched on a forest floor, connecting the device to a laptop before pushing it into the ground (Dunkelmann 2013). On the laptop screen, a matrix of flashing characters appears. Has the computer ‘booted’ or ‘crashed’? The earthboot device is designed to emulate a USB storage device, such as a memory stick or hard drive. However rather than reading stored data, the device samples electrical voltages from the earth while the computer starts up; these measured values are converted into instructions in a form that is legible for the host computer. They become opcodes and operands - low-level instructions for the host CPU. This translation process is relatively direct: “some values are discarded which are assigned to operations which don’t make much sense in this context but there is no great effort to understand the code the earth is writing semantically and see if this makes any sense to execute” (Howse, correspondence 2012). Not surprisingly, the earth’s code often appears in unfamiliar forms. As Howse deadpans: “Quite often the earthboot operating system is not always fully functional as expected. Crashing is the price to pay for booting straight from the earth” (2013a).
Earthboot creates a bridge or link between conventional digital computation and its earthly substrate. The device literally mediates between the electromagnetic currents of the earth, and the digital protocols of the booting computer; through a process of sampling, the earth is made (or enabled) to code. Howse presents the work as short-circuiting the mineral extraction and processing of conventional hardware; instead this device ‘extracts’ only data, measuring immanent differences within the earth. However like the proto-programming of Baecker’s Irrational Computing, data sourced from the material flux is converted here into code, executable instructions.
In Earthboot the ‘site of execution’ remains more or less where we would expect - within the digital machinery of a laptop. The work opens a vector of contagion, inviting an incursion of earth code to be executed by the host, just as a virus co-opts the procedures of the host cell. In his more speculative Earthcodes projects, Howse imagines more significant shifts in the nature and site of material computation. He describes the “earth computer” as
...a (computational) device of the same substance as the earth, to be embedded in the earth .... The central conceit is the use of the earth itself as a dirty, irrational computational device. An attempt will be made to reproduce common components, such as memory, power supply, and CPU with earth-based elements; a form of computational land art. Techniques borrowed from the semiconductor and computer industry will be applied to the raw earth substrate either in situ ... or as a speculative performance. (Howse 2013a)
Howse documents experiments in creating the earth computers’ components, including growing a ‘CPU’ by inoculating an oyster mushroom with semiconducting minerals such as galena, as well as compounds extracted (leached) from waste computer hardware. He proposes a ‘display’ made from piezoelectric crystals; tiny movements can be amplified into the perceptual domain using something like a mirror galvanometer, a mechanical device for measuring current dating from the late 19th century. ‘Memory’ could be made from ferrite cores: in the dominant technology for computer memory from the 1950s-1970s, this crystalline compound was formed into small toroidal ‘cores’ whose magnetic fields could be manipulated to store data (see Reilly 2003: 164). Power would come from an “earth battery” - a device transducing telluric magnetic currents into electricity, based on the 1898 invention of American Nathan Stubblefield (Stubblefield 1898).
The architecture here suggests that an earth computer might resemble more conventional devices, but in Howse’s final proposals this resemblance literally dissolves, and the notion of computation is radically expanded. Here the artist imagines a “machine without components”,
allowing the earth to manipulate and code computational/crystalline structures. Raw minerals, either dissolved or as solid plates are placed within a porous ceramic/glass structure to be buried within the earth .... Over time, it is anticipated that both underground electric currents (telluric flows) and minerals/rainwater leaching through the soil could re-form these base components (some extracted from computer waste) into a functioning earth computer .... (Howse 2013a)
Developing this model, Howse proposes Earthcode “monuments”, citing Robert Smithson’s land art and the unrealised monuments of Gustav Metzger (see for example Metzger 1971). Here, environmental processes “code or inscribe” the landscape - either via a robot excavator, ‘programmed’ by patterns of change in the surroundings, or more subtly, through the mediation of an in-situ earth computer, where earth batteries, leached and extracted minerals and ambient electromagnetic radiation combine. In such an “open” computer, Howse writes, “shafts of light penetrating the edges of the cleared area equally code this embedded landscape” (Howse 2013a).
Kittler’s “sheer hardware” imagines “maximal connectivity”; “a physical device working amidst physical devices”; a form of computing without software (Kittler 1995). Baecker and Howse implement this proposal, but they also reach far beyond it. In the process, they breach the conceptual models of computation that Kittler draws from Conrad, and move beyond a simple ‘grounding’ of computation in its material substrates, towards a more expansive reconfiguration of machine, matter and agency.
The physical immanence of “maximal connectivity” is evident throughout this work. In Baecker’s crystal field, physical and electromagnetic forces interact in a single dynamic plane; Howse’s earth computer experiments evoke immanent relations between mineral and living systems. Kittler writes of “maximising noise” and embracing physical phenomena that digital computing treats as limitations; this is clear in the chaotic signal generators of Baecker’s work in particular. Yet as Kittler draws on Conrad, he also imports a notion of computing that is fundamentally functional. The limitations of conventional digital computers that Kittler describes are limits to computational performance - a ceiling to the volume of information processed. “Sheer hardware” is essentially (following Conrad) a way to bypass those limits, to “keep up” with the complexity of the world or “enter that body of real numbers … known as chaos” (Kittler 1995). Kittler is elusive, but Conrad makes it clear that “keeping up” or “entering” here refers to representation and simulation; Conrad focuses in particular on “the capacity of programmable machines to simulate nature and duplicate intelligence” (Conrad 1995: 261). The underlying project here is the dream of perfect simulation, or the computability of the world.
Conrad sets out a formal model of machine, program and programmability, based on the Turing model of the computer as a symbol-manipulator. This model system involves a set of stored discrete symbols (the “tape”) manipulated one at a time according to a set of formal, deterministic rules (the “program”). For Conrad “programs are rules … that feature finiteness of the number of symbols employed, discrete differences among the symbols, and, in the action of the rule, discreteness in time” (1995: 261). Programmability is simply “the ability to prescriptively communicate a program to an actual system” (1995: 262).
It is here that the material computing of Baecker and Howse peels away from Conrad (and Turing). One of the key attributes of the Turing model is its dependence on discrete, symbolic elements; in any real computing system these symbols must be carefully stored on and transcribed from material substrates in a messy, analog world. The independence of symbol and substrate is a crucial attribute of the digital - Will Schrimshaw (2012) describes it aptly as an “indifference” to the material. Baecker and Howse show instead how the symbolic domain can be attuned to the flux of the material. In Baecker’s coincidence detector cosmic ray impacts are filtered from the environment using a proto-digital circuit; it emits a high or a low voltage, this or that, 1 or 0. This signal becomes a protean symbol as it switches or routes the signals of the other modules to the display unit. Similarly in the other Irrational modules, material chaos is sampled into discrete, regulated voltages, or coaxed into binary patterns, as in the shot noise generator, where the electro-mechanical interactions of wire and galena crystals form macroscopic diodes. Like a digital computer, Irrational Computing has a ‘clock’ - a signal dividing time into discrete steps, as per the Turing model: but here, that signal is a fluctuating pulse drawn from the crystal field oscillator; instead of being isolated and purified, the oscillations of quartz are mashed and multiplexed. In Howse’s Earthboot, the attunement of symbol to matter is even more explicit; here the electrical potentials of the earth are measured or sampled, then encoded in a digital form. Earth-signal becomes earth-symbol; and these symbols in turn are fed to a computing machine as its ‘program’. In a computer virus, a parcel of data (say a file attached to an email) tricks the host into interpreting it as code; this category error is facilitated by the interchangeability of code and data, their shared symbolic encoding. So too in Earthboot, telluric signals become data, then prospective code - a sort of earth-borne virus.
Howse and Baecker tear down the insulation between the symbolic, internal world of the computer, and the immanent material fluxes of substrate and environment. This also entails a reconfiguration of computational agency. In a Turing machine, agency is simple, linear and exclusive. The programmer devises a formal, explicit program; communicates it to the machine “proscriptively”, in Conrad’s words; and the machine executes the program, manipulating its own internal state. In the work of Baecker and Howse, distinctions between ‘internal’ and ‘external’ are quickly dissolved. To be attuned to matter is also to invite the outside in, and thus to give up exclusive agency over the machine’s symbolic representations. Earthboot demonstrates this very clearly, inviting earth-voltages in as CPU instructions - this incursion is a one-off ‘import’, and if the laptop boots, computation continues in familiar, insulated form. By comparison Irrational Computing is continuously open to its environment, switched (or programmed) by incoming cosmic rays. Howse’s speculative “earth computer” pushes this openness even further, catalysing earth-borne electrical currents, leaching and crystallisation, and imagining these as computations in a “machine without wires”. The unitary agency of the Turing model becomes ever more extensive, culminating in the immanent, self-programming earth computer. While Baecker’s architecture is relatively discrete and explicit, his work shares this model of immanent materiality. He cites the naturally occurring fission reactors discovered at Oklo, in Central Africa: concentrated uranium accumulated here through a long sequence of geological processes, triggering nuclear reactions (see Loss 2009). So technology is turned inside-out; and the purposeful, instrumental engineering of matter is only echoing an ancient, expansive, non-human agency.
These works also cultivate more intensive agencies. At every opportunity Baecker takes proto-digital components - crystal oscillators, diodes, logic gates - and wires them up to amplify their intrinsic instabilities. In the phase locked loop and the crystal field, generative chaos emerges firstly through the transductive capacities of the crystal materials: their two-way articulation of electrical and mechanical energy. Where traditional computing isolates and functionalises this agency - as in a quartz clock - Baecker multiplies it with itself, hooking it into amplifying cycles and networks of positive feedback. Notably in the shot noise generator, the same amplification of material instability occurs through a sort of attenuation: by reducing the current to a trickle, its microscopic fluctuations come to the fore. In the language of Deleuze and Guattari (1987: 272): the stable, “molar” identity of electrical current is revealed as a fiction, as the activity of its “molecular” constituents emerges.
While these material agencies certainly breach the Turing model of program and programmer, it would be simplistic to characterise this material computing as ‘unprogrammable’, or somehow superseding human agency. Certainly the artists here do not program their machines in the ‘proscriptive’ model that Conrad outlines; instead their ingenuity is collaborative and evocative, a working-with and drawing-out. They select, evoke, sample and amplify intensive and extensive agencies of earth and energy. Jane Bennett outlines a “distributive agency”: one “distributed across an ontologically heterogeneous field, rather than being a capacity localised in a human body or in … human efforts” (2009: 23). Bennett’s agency is a confederation, rather than a single locus of intent and execution: it arises within human-nonhuman “assemblages” characterised by an emergent causality. Baecker and Howse demonstrate this clearly, but they also go further; in reconfiguring computation as material, in pursuing ‘sheer hardware’, they also sketch the implications of a ‘distributive’ computational agency. In fact, their work suggests that any such materialised computing must involve a more complex model of agency than Turing’s program and programmer.
Kittler, Baecker and Howse are united in grounding computing in its material substrates. Kittler declares that “software, if it existed, would just be a billion dollar deal based on the cheapest elements on earth” (Kittler 1995). So too the artists direct us to the minerals and crystals on which the digital computer is founded, zooming up the tiny wafers of semiconducting silicon until they become brute rocks. Howse literally reinters computing, returning it to the earth. But as argued above, this rematerialisation goes much further than simply affirming the material substrate of computation. By tuning in to that substrate, rather than keeping it at a distance, Baecker and Howse show us Bennett’s “vibrant matter” - a generative, active material, but also a “material vitality” that we share with the world of living and nonliving things - what Bennett calls a “strange and incomplete commonality” (2009: 18).
As software “eats the world” or “takes command”, the question of how computing is - and how it could be - in the world, is timely. At the same time, our relations with the living and nonliving things around us need urgent rethinking; the “naive ambition” of Bennett’s materialism is just such a reconsideration (ibid). In keeping the world at bay - remaining ‘indifferent’ to matter - digital computing enacts a philosophical position, a normative configuration of subject, matter and agency. The works of Howse and Baecker hint at a rich and strange alternative.
Andreessen, M. (2011) “Why Software Is Eating The World” in Wall Street Journal, 20 August, viewed 30 July 2013, http://online.wsj.com/article/SB10001424053111903480904576512250915629460.html
Bennett, J. (2009) Vibrant Matter: A Political Ecology of Things, Durham: Duke University Press.
Blum, A. (2012) Tubes: Behind the Scenes at the Internet, London: Penguin UK.
Bonolis, L. (2011) “Walther Bothe and Bruno Rossi: The Birth and Development of Coincidence Methods in Cosmic-ray Physics” in American Journal of Physics 79: pp. 1133-1150.
Conrad, M. (1995) “The Price of Programmability” in The Universal Turing Machine A Half-Century Survey (Herken, R., ed.), Vienna: Springer. pp. 261-281.
Deleuze, G., and F. Guattari (1987). A Thousand Plateaus: Capitalism and Schizophrenia, Minneapolis: University of Minnesota Press.
Dunkelmann, M. (2013) “Die Erde ist Mein Motherboard” (Video recording), ARTECreative, viewed 25 November 2013, http://creative.arte.tv/de/magazin/martin-howse-die-erde-ist-mein-motherboard
Howse, M. (2013a) “The Earthcodes Project: Substrate/shifting the Site of Execution”, viewed 1 August 2013, http://www.1010.co.uk/org/earthcode.html
Howse, M. (2013b) “_____ -micro Research [London/Berlin]”, viewed 1 August 2013, http://www.1010.co.uk/org/
Kirschenbaum, M. (2005) “‘Every Contact Leaves a Trace’: Computers, Forensics and Electronic Textuality” in Workshop in the History of Material Texts, University of Maryland, vol. 4.
Kittler, F. (1995) “There Is No Software” in CTheory 10 http://www.ctheory.net/articles.aspx?id=74
Loss, R. (2009) “The Oklo Fossil Fission Reactors”, viewed 1 August 2013, http://oklo.curtin.edu.au/
Maeda, J. (2004) Creative Code, London: Thames & Hudson.
Manovich, L. (2013) Software Takes Command, London: Bloomsbury Academic.
Oliver, J., Savičić, G. & Vasiliev, D. (2011) “The Critical Engineering Manifesto”, viewed 31 July 2013, http://criticalengineering.org/
Parikka, J. (2012) “New Materialism as Media Theory: Medianatures and Dirty Matter” in Communication and Critical/Cultural Studies 9 pp. 95-100.
Reilly, E. (2003) Milestones in Computer Science and Information Technology, Westport: Greenwood Publishing Group.
Schrimshaw, W. (2012) “Undermining Media” in Artnodes 0: 12, http://in3wps.uoc.edu/index.php/artnodes/article/view/n12-schrimshaw.
Stubblefield, N. (1898) “Electrical Battery”, US Patent US600457 A.
Whitelaw, M. (2011) “After the Screen: Array Aesthetics and Transmateriality” in Column 7 pp. 50-57.
Whitelaw, M. (2012) “Transmateriality: Presence Aesthetics and the Media Arts” in Throughout: Art and Culture Emerging With Ubiquitous Computing (Ekman, U. ed.), Cambridge: MIT Press. pp. 223-236.
Ralf Baecker, “Irrational Computing”, 2011. http://www.rlfbckr.org/work/irrational_computing
Ralf Baecker, “Irrational Computing / Documentation”, 2011. http://vimeo.com/37443273
Gustav Metzger, “Projects Unrealised 1”, 1971. http://www.tate.org.uk/art/artworks/metzger-projects-unrealised-1-t12338
Mitchell Whitelaw is an academic, writer and practitioner with interests in new media art and digital culture. His writing has appeared in journals including Leonardo, Digital Creativity, Fibreculture, and Senses and Society. His work on rich interfaces to digital cultural collections has been supported by the State Library of New South Wales, the National Archives of Australia, the National Museum of Australia, and the National Gallery of Australia. He is currently an Associate Professor in the Faculty of Arts and Design at the University of Canberra, where he leads the Digital Treasures PhD program.