All posts by Annet Dekker

The Question of Documents

by Annet Dekker

— short intro: the following is a draft version and part of my last chapter in my thesis, hence some obvious information or links may be missing. in my presentation i will concentrate on the analysis of several case studies that underlie this more theoretical part. also note that it hasn’t been proofed for english. —

In her seminal text Qu’est-ce que la documentation (1951) Suzanne Briet expanded the notion of the document to also include natural objects and works of arts. Documents were regarded as examples or grouping of things, which derive meaning from their context. This approach is still valid today but it would need to be redefined and clarified, because what happens when the context, for example a distributed network, is the work? Are software and algorithms also documents? Is something immaterial, a process, or a network a document – and if they are not, then what are they?

Similar to Briet, Lev Manovich argues that it is not enough to examine the ‘final’ presentation in order to understand contemporary media; social, historical and technological contexts should be taking into consideration when talking about or identifying documents. However, Manovich uses the term ‘software performances’ instead of documents because ‘it is software which defines the options for navigating, editing, and sharing the document, rather than the document itself’ (2013:34), thereby stressing the construction by software of experiences. The discussion of whether the term document is still useful in a digital age is also brought up by David Levy (1994 and 2001), and others like Michael Buckland (1998), although they don’t come up with a solution, both argue to follow the path of the earlier documentalists (among others Paul Otlet and Suzanne Briet) by focusing on defining a document in terms of function rather than physical format. Although it is striking that Manovich doesn’t refer to documentalists practices, his descriptions and analyses follows a similar approach of trying to answer the question of what constitutes a ‘document’, or in Manovich’s terms, to understand media software. So, in what way is the notion of ‘software performances’ useful, and should it replace the term document? What does performance mean in relation to software? Which aspects perform? For what purpose? For whom?

One of the main characteristics that I use to describe net art is its performative qualities. Net art can be understood as performative in terms of the meanings ascribed to it as well as in terms of the effects of its performance on the movements of data and information in communication networks. The verb ‘perform’ means to act, to carry out an action or pattern of behaviour.[i] In the context of art, perform or the noun performance, is mostly associated with Performance art. Although the term Performance art is a contested concept (Carlson 1996), in general it refers to a performance presented to an audience in which the performer(s) doesn’t present a conventional theatrical play or a formal linear narrative. Phillip Auslander emphasises that in traditional terms it may be problematic to see bots (or technical tools in general) as performers, because such definitions generally emphasize the performer as someone who executes and in that process makes interpretations that lead to specific aesthetic effects.[ii] In order to make his argument he makes the distinction between technical and interpretive skills. When analyzing the installation Listening Post (2002 -) by Mark Hansen and Ben Rubin, Auslander argues that the installation is an example of technical performativity, because it ‘constructs its performances by sampling [live] conversations on the Internet’ (2005:8). Auslander continues that ‘the particular technical skills possessed by Listening Post could not be found in a human performer, for no human being could scour the Internet, gather data, sort it, and display it in real time with the speed and accuracy of the machine’, thereby stressing the speed and accuracy of the technical skills of the computer. The use of digital artworks as examples of performance art and in performance studies is becoming more common.[iii] However the distinction between the technical and interpretative skills is supported in most cases. Although unarguably computers are incapable of human interpretation in the sense of reading between the lines or making assumptions, I’d like to argue that software programmes, especially in algorithmic processes, can perform in complex ways that go beyond a technical narrative as emphasized by Auslander. Such ‘performativity’ enacts what it represents or describes, furthermore connecting performativity with ‘cultures of circulation’, as discussed by Benjamin Lee and Edward LiPuma (2002), opens the discussion to see software performances as creators of the act they refer to. Finally these arguments will challenge the meaning of the term document.

The term performativity derives from British philosopher of language J.L. Austin. In his publication How to Do Things With Words (1962) he describes performative utterances as statements that perform an action: a Speech Act. Rather than describe or report what is being done, they do (1962:5).[iv] At first sight Austin’s Speech Act theory fits the model of computation, which generally breaks down in three stages: input, processing and output.[v] An input into the system does something, physically in the voltages and in the mechanisms of the machine, and computationally in the abstract mathematics of processing.[vi] But, similar as Austin’s theory simplifies the context of language and meaning by regarding it as a ‘total situation’ (1962:52), as also emphasized by Derrida (1988[1972]), there is always uncertainty and ambiguity present in processing. As Arnold Michelson and Allen Levesque argue ‘It is clear from the outset that with any real communication system we cannot expect to receive exactly what is transmitted. At the very least, we can expect noise to be added to the transmission, causing random errors’ (1985:4).[vii] Moreover, leaning on Claude Shannon’s communication model, Susan Ballard explains that information cannot occur when there is no noise in the process (2007).[viii] This means that performativity has always a certain level of unpredictability, uncertainty and ambiguity, or in other words that the input and output are not necessarily coherent.[ix]

Performativity is used by many artists, either actively as in the case of The English performance group Blast Theory by making failing hard- and or software part of the overall performance, or as artifacts of historical instances in the case of Martine Neddam’s (1997 –) by holding on to some errors instead of fixing them.[x] Such performativity of code means that code is not one-to-one reversible, nor can it be seen as pre-set instructions for execution. Performativity of code indicates that execution takes place by thinking through the material. As such, the challenge lies more in the question whether code or software performances also create the ‘act it refers to’. The ‘act’ of is often associated with identity play, Blast Theory with game adventures and Naked on Pluto with addressing privacy issues, but in what way do they also act beyond these meaning-making narratives? How do they engage in and facilitate circulation, one of the main characteristics of net art? Would a focus on circulation and process offer a means to critically address the performativity of net art? In their article ‘Cultures of Circulation: The Imaginations of Modernity’ (2002) Lee and LiPuma propose an alternative version of the concept of performativity. They see performativity as an aspect of circulation, rather than as a central concept of meaning-making. ‘Performativity has been considered a quintessentially cultural phenomenon that is tied to the creation of meaning, whereas circulation and exchange have been seen as processes that transmit meanings, rather than as constitutive acts in themselves’ (2002:192). They continue: ‘Cultures of circulation are created and animated by the cultural forms that circulate through them, including—critically—the abstract nature of the forms that underwrite and propel the process of circulation itself’ (2002:193).[xi] In net art such circulation can be traced by looking at how movement performs in the code, in the interaction between code, programmer and context and how this shapes visitors’ experiences. As well, such circulation and exchange of code involved in the infrastructure of communication may reveal specific power structures.[xii]

Document as process, or process as document

How could the concept of performative circulatory and processes help with the conservation of net art? My emphasis on the processual dimensions of materiality suggests that what something is has to be understood in terms of what it does, how it (historically) works within machinic, systemic, social and cultural domains. In order to understand and critically reflect on the evolution and the political dimensions inherent in computation it is important to study these processes, their behaviour, how they function, and how they are embedded in and influenced by social and technical contexts. Such an approach will also guide conservators to answer what the material is, what the intention of the artist(s) was, and find ways to capture, restore or document net art.

With processes being the work, or seeing the work as a process, Renee van de Vall suggests to speak of a third paradigm in conservation, the first being centered around scientific conservation (or the autographic paradigm) and the second, leaning on Pip Laurenson (2005) around performance and performative behaviour.[xiii] In her view this third processual paradigm can be characterised by artworks that are following ‘rules of the game’, are open-ended, in continuous development, and part of the development of the work is outsourced (either by technical or natural processes, or participants). These artworks unlike performative works are not predefined by instructions or notations and as such Van de Vall makes an analogy with improvised music, and stresses that it is not a matter of one paradigm substituting for the other, but that these approaches can be seen to work in parallel and even at times intermingle. My findings seem to support the division between performative and processual artworks. However, it remains to be seen if such a clear separation is necessary when discussion conservation, or documentation strategies. For example are ‘rules of the games’ (sending something out into the world and let it evolve) the same as a ‘set of instructions’ (there is a margin of variability but not everything goes).[xiv] In most cases there will always be some kind of restriction,[xv] either through the set up of the artwork, for example in the case of, most parts of are still linked or kept together by the main website, and the participants are encouraged to remain within the domain. The game-engine of Naked on Pluto acts in such a way that it is processual, because the game-engine is generative, but the game itself is only partly so, the rules of the game are pretty fixed and not everything goes. The performances of Slub World are probably the closest to the characteristics of a processual paradigm, but even here it could be argued that it is not only generative. Even though it is based on algorithmic processes and that process is the narrative, the human input is very important.[xvi] As McLean describes:

‘In live coding the performance is the process of software development, rather than its outcome. The work is not generated by a finished program, but through its journey of development from nothing to a complex algorithm, generating continuously changing musical or visual form along the way’ (McLean 2011:130).

Looking at degradation of fabrics or other biological material are these also processual? Such works evolve, but there is no sense of (outsourced) participation, with some variations they can be brought back to their ‘original’ state. To sum up, the logics in most works can be analytically different, which is important for understanding and analysing a work, but most artworks have performative and processual elements. So, what could be the consequences for conservation, will these two paradigms need different approaches or strategies? It seems obvious that with processual works conservation in the strict sense will not be possible, but the same could be argued for many performative artworks. In both cases documentation will likely play a more important role than the reconstruction of the artwork.



Auslander, Philip (2005) At the Listening Post, or, do machines perform? International Journal of Performance Arts and Digital Media. Vol.1, nr.1, pp. 5-10.

Ballard, Susan (2007) Information, Noise and et al. M/C Journal, Vol.10, Issue 5.

Briet, Suzanne (2006, originally published 1951) What is documentation? (Qu’est-ce que la documentation?) Translated and edited by Ronald E. Day and Laurent Martinet with Hermina G.B. Anghelescu. Lanham, MC: Scarecrow Press.

Buckland, Michael (1998) What is a ‘Digital Document’? Document Numérique (Paris) Vol.2, No.2, pp. 221-230.

Carlson, Marvin (1996) Performance: A Critical Introduction. London and New York: Routledge.

Derrida, Jacques (1988, English translation [1977] from French [1972]) Limited Inc. Evanston, IL: Northwestern University Press.

Laurenson, Pip (2006) Authenticity, change and loss in the conservation of time-based media installations. Tate Papers, Issue 6.

Lee, Benjamin and Edward LiPuma (2002) Cultures of Circulation: The Imaginations of Modernity. Public Culture. Vol.14, No.1, pp. 191-213.

Levy, David M. (1994) Fixed or Fluid? Document Stability and New Media. European Conference on Hypertext Technology 1994 Proceedings. New York: Association for Computing Machinery, pp. 24-31.

Mackenzie, Adrian (2005) The Performativity of Code: Software and Cultures of Circulation. Theory, Culture & Society, Vol.22, No.1, pp. 71-92.

Manovich, Lev (2013) Software Takes Command. Cambridge, MA: The MIT Press.

McLean, Alex (2011) Artist-Programmers and Programming Languages for the Arts. Ph.D. thesis, Department of Computing, Goldsmiths, University of London.

Michelson, Arnold M. and Allen H. Levesque (1985) Error-Control Techniques for Digital Michigan: The University of Michigan Press.


[ii] Auslander bases his argument on the quote from philosopher Stan Godlovitch who discusses musical performance, ‘interpretive skills involve aesthetic effects for which no obvious quantitative measure exists, and typically emphasize “expression” …’(Godlovitch 1998:54, in Auslander 2005:6).

[iii] See among others, Bay-Cheng, (2010), Giannachi, (2012), Bleeker (2013).

[iv] Austin further distinguishes between an ‘illocutionary act’ that is concerned with what someone/something is doing when saying something, and a ‘perlocutionary act’ that involves the consequence(s) of an utterance. The utterance and the consequences of that utterance don’t occur at the same time. According to Austin in order for the illocutionary act to be successful certain conditions need to be met. However, as pointed out by Derrida (1977[1972]) meaning, nor context, of a text cannot be defined in its entirety – a performative utterance is always intertwined with structures of power.

[v] See among others Charles Petzold (2000) and Ive Englander (1996).

[vi] It may be good to stress that I’m referring here to formal executions; it is not a social performance based on human conventions (as in Austin’s theory). Technologies, in and of themselves, do not bring about cultural or social change.

[vii] In modern computers many processes and redundancies are build in to reduce the effects of noise, making it unlikely that a computational error will occur. Nevertheless the more complex processes become, the more noise comes in which can lead to unexpected or unnoticed events. However even in ‘simple’ systems, like CRT and LED monitors ‘single transmitted voltage might simultaneously perform the one or zero of binary code, disrupt adjacent data with its electromagnetic noise, and be received as radio waves by an external antenna’ (Van Orden 2010), a process that was named Van Eck Phreaking. See also Van Eck (1985) and Kuhn (2004).

[viii] Many artists have used these errors (also referred to as Glitch) to make artwork, for more information see among others Goriunova and Shulgin (2008) and Menkman (2011).

[ix] Live coders explore these characteristics of programming in their live performances. ‘Live coding is the activity of writing (parts of) a program while it runs. It thus deeply connects algorithmic causality with the perceived outcome and by deconstructing the idea of the temporal dichotomy of tool and product it allows code to be brought into play as an artistic process’ (Alexander, 2004:243-244 ). See Yuill (2008) for more information on a historical contextualising of code practices referencing scratch orchestra of the 1960s.

[x] There are also other examples when ambiguity through performativity takes place, for example in the before mentioned use of identity in However, such performativity refers more to the meaning ascribed to performativity.

[xi]  In network theories the new forms of access, understanding and engagement with circulatory networks are explored (Benkler 2007; Castells 1996; Wittel 2001), but little attention has been paid to the dynamics of circulation itself as force of change.

[xii] I’m leaning here on the article ‘The Performativity of Code: Software and Cultures of Circulation’ by Mackenzie (2005) in which he asserts that ‘if we accept that information and communication constitute a central venue for the performativity of some important contemporary forms of power, then the circulation and exchange of software and code involved in the infrastructure of communication could well be analysed in performative terms’.

[xiii] Renee van de Vall (2013) ‘Documenting Dilemmas. On the Relevance of Ethically Ambiguous Cases’, keynote lecture at Performing Documentation in the Conservation of Contemporary Art, Lisbon 20-21 June.

[xiv] The term instructions is used by Laurenson to describe performative artworks, following Stephen Davies she argues that a ‘notation has the function of specifying works. A score is intended as instructions to potential performers and ‘it is by following these instructions that players generate instances of the work’’ (2005).

[xv] Generative artworks can be seen as the exception, for more information see (accessed 9 August 2013).

[xvi] What exactly defines generative art is still being discussed, most of these discussions centre around the human influence on the programme. See among others, Galanter (2003) and Mclean (2011:16-17, 115-127).