Information, Noise and et al.
Susan P. Ballard · M/C Journal · 2007
The two companions scurry off when they hear a noise at the door. It was only a noise, but it was also a message, a bit of information producing panic: an interruption, a corruption, a rupture of communication. Was the noise really a message? Wasn’t it, rather, static, a parasite? Michael Serres, 1982. Since, ordinarily, channels have a certain amount of noise, and therefore a finite capacity, exact transmission is impossible. Claude Shannon, 1948. Reading Information At their most simplistic, there are two means for shifting information around – analogue and digital. Analogue movement depends on analogy to perform computations; it is continuous and the relationships between numbers are keyed as a continuous ordinal set. The digital set is discrete; moving one finger at a time results in a one-to-one correspondence. Nevertheless, analogue and digital are like the two companions in Serres’ tale. Each suffers the relationship of noise to information as internal rupture and external interference. In their examination of historical constructions of information, Hobart and Schiffman locate the noise of the analogue within its physical materials; they write, “All analogue machines harbour a certain amount of vagueness, known technically as ‘noise’. Which describes the disturbing influences of the machine’s physical materials on its calculations” (208). These “certain amounts of vagueness” are essential to Claude Shannon’s articulation of a theory for information transfer that forms the basis for this paper. In transforming the structures and materials through which it travels, information has left its traces in digital art installation. These traces are located in installation’s systems, structures and materials. The usefulness of information theory as a tool to understand these relationships has until recently been overlooked by a tradition of media art history that has grouped artworks according to the properties of the artwork and/or tied them into the histories of representation and perception in art theory. Throughout this essay I use the productive dual positioning of noise and information to address the errors and impurity inherent within the viewing experiences of digital installation. Information and Noise It is not hard to see why the fractured spaces of digital installation are haunted by histories of information science. In his 1948 essay “The Mathematical Theory of Communication” Claude Shannon developed a new model for communications technologies that articulated informational feedback processes. Discussions of information transmission through phone lines were occurring alongside the development of technology capable of computing multiple discrete and variable packets of information: that is, the digital computer. And, like art, information science remains concerned with the material spaces of transmission – whether conceptual, social or critical. In the context of art something is made to be seen, understood, viewed, or presented as a series of relationships that might be established between individuals, groups, environments, and sensations. Understood this way art is an aesthetic relationship between differing material bodies, images, representations, and spaces. It is an event. Shannon was adamant that information must not be confused with meaning. To increase efficiency he insisted that the message be separated from its components; in particular, those aspects that were predictable were not to be considered information (Hansen 79). The problem that Shannon had to contend with was noise. Unwanted and disruptive, noise became symbolic of the struggle to control the growth of systems. The more complex the system, the more noise needed to be addressed. Noise is both the material from which information is constructed, as well as being the matter which information resists. Weaver (Shannon’s first commentator) writes: In the process of being transmitted, it is unfortunately characteristic that certain things are added to the signal which were not intended by the information source. These unwanted additions may be distortions of sound (in telephony, for example) or static (in radio), or distortions in shape or shading of picture (television), or errors in transmission (telegraphy or facsimile), etc. All of these changes in the transmitted signal are called noise. (4). To enable more efficient message transmission, Shannon designed systems that repressed as much noise as possible, while also acknowledging that without some noise information could not be transmitted. Shannon’s conception of information meant that information would not change if the context changed. This was crucial if a general theory of information transmission was to be plausible and meant that a methodology for noise management could be foregrounded (Pask 123). Without meaning, information became a quantity, a yes or no decision, that Shannon called a “bit” (1). Shannon’s emphasis on separating signal or message from both predicability and external noise appeared to give information an identity where it could float free of a material substance and be treated independently of context. However, for this to occur information would have to become fixed and understood as an entity. Shannon went to pains to demonstrate that the separation of meaning and information was actually to enable the reverse. A fluidity of information and the possibilities for encoding it would mean that information, although measurable, did not have a finite form. Tied into the paradox of this equation is the crucial role of noise or error. In Shannon’s communication model information is not only complicit with noise; it is totally dependant upon it for understanding. Without noise, either encoded within the original message or present from sources outside the channel, information cannot get through. The model of sender-encoder-channel-signal (message)-decoder-receiver that Shannon constructed has an arrow inserting noise. Visually and schematically this noise is a disruption pointing up and inserting itself in the nice clean lines of the message. This does not mean that noise was a last minute consideration; rather noise was the very thing Shannon was working with (and against). It is present in every image we have of information. A source, message, transmitter, receiver and their attendant noises are all material infrastructures that serve to contextualise the information they transmit, receive, and disrupt. Figure 1. Claude Shannon “The Mathematical Theory of Communication” 1948. In his analytical discussion of the diagram, Shannon actually locates noise in two crucial places. The first position accorded noise is external, marked by the arrow that demonstrates how noise is introduced to the message channel whilst in transit. External noise confuses the purity of the message whilst equivocally adding new information. External noise has a particular materiality and enters the equation as unexplained variation and random error. This is disruptive presence rather than entropic coded pattern. Shannon offers this equivocal definition of noise to be everything that is outside the linear model of sender-channel-receiver; hence, anything can be noise if it enters a channel where it is unwelcome. Secondly, noise was defined as unpredictability or entropy found and encoded within the message itself. This for Shannon was an essential and, in some ways, positive role. Entropic forces invited continual reorganisation and (when engaging the laws of redundancy) assisted with the removal of repetition enabling faster message transmission (Shannon 48). Weaver calls this shifting relationship between entropy and message “equivocation” (11). Weaver identified equivocation as central to the manner in which noise and information operated. A process of equivocation identified the receiver’s knowledge. For Shannon, a process of equivocation mediated between useful information and noise, as both were “measured in the same units” (Hayles, Chaos 55). To eliminate noise completely is to sacrifice information. Information understood in this way is also about relationships between differing material bodies, representations, and spaces, connected together for the purposes of transmission. It, like the artwork, is an event. This would appear to suggest a correlation between information transmission and viewing in galleries. Far from it. Although, the contemporary information channel is essentially a tube with fixed walls, (it is still constrained by physical properties, bandwidth and so on) and despite the implicit spatialisation of information models, I am not proposing a direct correlation between information channels and installation spaces. This is because I am not interested in ‘reading’ the information of either environment. What I am suggesting is that both environments share this material of noise. Noise is present in four places. Firstly noise is within the media errors of transmission, and secondly, it is within the media of the installation, (neither of which are one way flows). Thirdly, the viewer or listener introduces noise as interference, and lastly, it is present in the very materials thorough which it travels. Noise layered on noise. Redundancy and Modulation So far in this paper I have discussed the relationship of information to noise. For the remainder, I want to address some particular processes or manifestations of noise in New Zealand artists’ collective, et al.’s maintenance of social solidarity–instance 5 (2006, exhibited as part of the SCAPE Biennal of Art in Public Space, Christchurch Art Gallery). The installation occupies a small alcove that is partially blocked by a military-style portable table stacked with newspapers. Inside the space are three grey wooden chairs, some headphones, and a modified data projection of Google Earth. It is not immediately clear if the viewer i