top of page

Between the Archive and the Anarchivable - by Wolfgang Ernst

N°1 – Interrupting the Anarchival discourse

From the archive to the anarchival impulse

and back again

In academic, cultural and aesthetic discourses on storage and memory, a central epistemological focus has emerged: the "archival impulse", manifested in an "idiosyncratic probing into particular figures, objects, and events" (1) - even if this does not relate really to the archive but to its playful simulation, mistaking archival aura for the rigid power of the symbolic order. In more adventurous avantgardistic manifestos, this archive has recently been counter-balanced by a celebration of the "anarchive". The anarchival here serves "as a form of counter-knowledge production, as a dynamic that unlocks, liberates the archive". (2) This artistic and theoretical gesture of liberation (including some of my own previous writings (3) has developed a momentum of its own which now asks for flipping back, since it would be media-politically naive to confuse the current hypertextual World Wide Web with its underlying techno-mathematical substructures of algorithms embedded in the storage-programmable computer and its literally dynamic Random Access Memory (DRAM) which are more strictly "archival" than ever in its media-archaeological realities. Maybe digital media art tends to the anarchival memory aesthetics(4); the digital media archive does not. Rather than submitting to the political phantasies of anarchic memory politics, let us remember that a most rigid archival order is still the basis for digital memory and communication technologies.


The (still) tangible digital archive: Magnetic Core Memory (RAM) from the first generation of electronic computers. The archive is present both as content (each core one bit) and as address structure (the x/y grid). Photograph (Ines Liszko) from the Media Archaeological Fund at Media Studies, Humboldt University, Berlin.

In between the archival and the anarchival, the dynarchive has emerged. Digital data need constant up-dating (in terms of software) and "migration" (in terms of hardware to embody them). From that derives a change from the ideal of archival eternity to permanent change. Both the archive in its media base and the archive as discourse have literally got in motion, as is indicated by terms like the "processual" archive. (5) By definition, the "storage-programmable" von-Neumann-architecture of current digital computers interlaces real-time processing and intermediary storage of data. Micro-archiving the present has become the signature of digital culture in times of online communication media. Conceptually, the dynamic archive is relational not only in a structural or spatial but as well in a temporal sense. But how can the concept of the archive be opened to heterochronic experimentation and at the same time fulfil its traditional task of keeping a well-defined order intact for transmission into future memory? In classic taxonomy (especially in the Prussian state tradition), the primary task of the archivist is to keep the configuration of incoming record groups intact. This clashes with the data bank aesthetics of the digital age which is characterized by the mutability not only of the archival records themselves but of its archival infrastructure (both hardware and software) as well. Thereby the traditionally enduring time base of the archive itself is replaced with restless reconfiguration. Archival endurance (oscillating between historical and entropical time) is undermined when a record is not fixed any more on a permanent material storage medium like parchment but takes places electronically; flow (the current) replaces the static inscription. Michel Foucault's Archéologie du Savoir (1969) once translated the Kantean a priori of sensation (space and time) into his definition of l'archive as the laws of what can be expressed and administrated at all. "It is also possible, however, to understand technological a prioris in a technological sense." (6) Media archaeology understands the archive and the "archival drive" (Mackenzie) in its Kantean and Foucauldian sense: as the a priori of the techno-logical event, the condition of possibility for electronic signals and data to circulate and be retrieved at all, "generated by the referencing and storage structures of the networks themselves". (7) The Internet oscillates between the archive and the anarchive, depending on the media-archaeological or phenomenological point of view. (8)

Towards an informational aesthetics of archival value: (Neg-) Entropy and dis/order

in times of binary information processing


Radical media archaeology discovers the informational value from within the objects stored in an archive, such as the histograms in digital image processing, calculating the aesthetic entropy of an image. The economic concept of chaotic store administration corresponds with this hashing approach. Let us therefore redefine archival value in terms of information theory. The current fascination with the "anarchival" as discursive category corresponds with a functional core criterium of techno-mathematical communication theory: the signal-to-noise ratio. Twentieth century communication engineering has resulted in a positive connotation of what culturally had been rejected for centuries: noise. In a parallel sense, disorder (from the point of view of second order cybernetics) has become a situation not to be afraid of any more. Mathematical statistics and stochastics has been developed in the nineteenth century to cope with death rates in life insurance policies (on the level of social administration) and with the laws of thermodynamics in energy management. According to the Second Law of Thermodynamics each system tends, when mapped upon the temporal axis, to increasing dis-order. Ludwig Boltzmann's calculus of entropy (the tendency from order to disorder as a physical manifestation of the arrow of time) has been used as a model for measuring the degree of probabilities in digital information theory. In terms of techno-mathematical communication theory, archival value looses its apparent semantic meaning in favour of statistical probabilities. It may be written or spoken words, pictures or music: The information source in this model "selects a desired message out of a set of possible messages" (10) - a virtual archive, with the notion of the archive itself turned upside down. This concept of information applies not to the individual meaningful message but rather to the situation as a whole - order wrenched from disorder. (11) Archives are indeed not simply storage as time channel but primarily defined by their records filtering function which is a quality that automated search engines mostly lack. Archival value creation such as algorithmic data mining can thus be claimed as negative entropy. Whereas statistics is still an "archival" (list-based) approach, stochastics (deciphered as Markov chains) shifts the past / present correlation to the present / future by predictive analytics.

N°2 – "Sound" and the Un–Archivable (12)

A different kind of recording: The

phonographic un-archive

Is there something like a sono-cultural archive? Sonic memory still tends to be neglected in visual culture studies. Previous to phonographic recording, voice and sound have escaped the archive, since it is rather un-archival in its temporal essence; it is radically signal-based, not symbolically ordered by alphabetic writing and musical score notation. Sound and speech have been the most “immaterial” cultural articulation (before the electronic age) so far. Let us therefore apply a Derridean différance, a significant shift which makes sense only in alphabetic scripture rather than phonetic enunciation: the "unarchive" instead of the "anarchive". The archival regime traditionally consists of alphabetic symbols both in its content (textual records) and is administration in meta-data (the inventory). But with technical audio signal registration, a different kind of recording emerged with does not belong to the symbolical regime any more - until with binary digitization the "Gutenberg Galaxy" re(oc)curred again. The phonograph with its simple metal needle permits the recording of vibrations that human ears previously could not memorize and writing hands never caught up with. The continuous undulations recorded by the gramophone and the audiotape keep the signature of the real, or physically raw audio material, replacing literature by sound engineering. (13) The phonographically recorded acoustic real "forms the waste or residue that neither the mirror of the imaginary nor the grid of the symbolic can catch: the physiological accidents and stochastic disorder of bodies". (14) The BBC World Service once launched the "Save our Sounds" project, looking to "archivize" sounds that may soon be lost due to the post-industrical world. (15) But different from the archive as symbolic order composed of records in historiographic, that is in alphabetic notation, such subtextual, signal-based recording is an unarchival modality: the sound of times past. So far historiography has privileged the readable archival records (in accordance with McLuhan's diagnosis of the Gutenberg Galaxy being dominated by visual knowledge). But since Edison's phonograph, sound, noise and voices can be technically recorded and thus memorized. The phonograph (respectively Emile Berliner's grammophone) registers an impressive range of acoustic events. Whereas musical notation (developed by the Greeks and Guido of Arezzo in analogy to the alphabet) is still symbolic recording, the phonograph registers the physically real signal. While alphabetic symbolism reduces acoustic events to the "musical" (harmonic) order, the register of the acoustic real encompasses the whole range of the sonic, including noise and arhythmical temporal phase shifting such as "swing" – non-archival sound in analog storage as opposed to the archival order of musical notation. As a research method, an archaeology of listening to past sound differs from Hearing History (16); it does not simply aim to widen the range of source material for historians. Let us propose something like a "diagrammatic listening", with the diagram having no indexical or iconic relation to a concrete sound, but rather represents a structural sound-image (German Hörbild). So-called Humanities have for the longest time not been concerned with the physically real – due to the limits of hermeneutics as text-oriented method, to the privileging of narrative as dominant form of representation and because of an essential lack of non-symbolic recording media of the real. Battles have been described and interpreted, but the real noise and smell of a combat could not be recorded for the archive and transmitted until the arrival of the Edison phonograph. (17) Phonography did not just provide historical research with a new kind of source material; it rather articulated new, rather ahistorical forms of tempor(e)ality on the level of the physically and mathematically real (which is, literally, techno-logy). With the phonograph, hearing became attentive of all kinds of sounds, regardless of their source, quality and meaning(lessness), just like the inner ear impassionately transduces vibrations analogue to electro-mechanical sound reproduction. With phonographic recording, listening became ahistorical, subject to the time-invariant reproducibility of acoustic signals. The phonograph provided sound production with a different kind of unhistorical index (in terms of Walter Benjamin), preserving its unique "auratic" experience, keeping the aural quality (in both senses) of its time quality - since a tone exists only in transience, that is: as Husserlean "time-object". (18) The Greek term mousiké encompassed not only musical articulation in the narrow sense, but dance and poetry as well. The essential operation to create a memory of such "time-based" arts, of course, is recording: either as symbolical archive (by dance notation in the tradition of writing as graphé), or by media endowed with the capacity to un-archivally register the physically real audiovisual signals.

Remembering past sonospheres by

technical media

The earliest effort of sound recording was a symbolic machine: language written in the phonetic alphabet. But the presence-generating power of technically recorded voices differs fundamentally from the grama-phonic notation of speech in the vocal alphabet. With the refinement of the Phenician alphabet to the Greek phonetic alphabet (which Ong actually called a "technologizing of the word" (19), acoustic articulation (speech, singing, oral poetry) became symbolically recordable for re-play. But symbolic notation (musical scores end in paper archives) differs from signal recording (which rather result in media collections):

"Musical notation [...] is about eternity: it kills music as a natural phenomenon in order to conserve it — once it is broken — as a spiritual entity: The survival of music in its persistence presupposes the killing of its here and now [...]."(20)

Edison cylinders need archival protection, for sure. But once Caruso's voice is articulated by re-playing its recording on Edison cylinder, human perception forgets about the archive. Historians among us will probably responded with reservations on this point, stressing that there is no unmediated access to the past. But let us not forget: Technologies of memory, addressing our perception on the affective rather than cognitive level, successfully dissimulate the archive. "Discourse analysis cannot be applied to sound archives or towers of film rolls." (21) But there is another, rather un-archival type of sound and noise sources - which is close to what humanities and classical studies know since long as the archaeological object. "Hearing the cracks and noises of a phonograph recording may initially enlighten their historical status as 'mechanical' instruments." (22) In terms of the mathematical theory of communication (Shannon 1948), such cracks belong to the kind of "noise" introduced by the channel of transmission itself which is here: the channel called time. When listening to ancient recordings from Edison wax cylinders, nowadays being restored with technomathematical software as digital re-production of sound, we might ask with Michel Foucault (in a slightly different context (23): Message or noise? To perform a psycho-acoustic experiment of a very simple kind, let us imagine an ancient phonographic recording of a song or voice. Whatever the timbre might seem, one will acoustically hallucinate as well the scratching, the noise of the recording apparatus. True media archaeology starts here: The phonograph as media artefact does not only preserve the memory of cultural semantics but "archivizes" past technical knowledge as well, a kind of frozen media knowledge embodied in engineering and waiting to be un-revealed by media-archaeological consciousness. Listening to technology really means close listening to the technological artefact itself. The Museum of Endangered Sounds takes care of the sound of "dead media" (24), and the Technical Committee of IASA in its recommendations from December 2005 insists that the originally intended signal is just one part of an archival audio record; accidental artefacts like noise and distortion are part of it as well - be it because of faults in the recording process itself or as a result of later damage caused in transmission. Both kind of signals, the semantic and the Proustean mémoire involontaire, message and noise, need to be preserved in media-archival conservation ethics. Media-archaeological listening to the sonic past is rather about the technical signifier than the musical signified. When we listen to an ancient phonographic record, the audible past (25) very often refers rather to the noise of the recording device (the ancient wax cylinder) than the recorded voice or music. Here, the medium talks both on the level of enunciation and of reference. What do we hear most: the cultural content (the formerly recorded songs) or the medium massage such as limitations in vocal bandwidth, even noise (the wax cylinder scratch and groove)? With digital sampling and processing of audio-signals, noise resulting from the frictions of analog technologies is usually significantly filtered, thus: silenced. But the former noise is being replaced by an even more endangering challenge: the "quantizing noise" on the very bit-critical (technical) level of signal sampling, and the migration problems of digital media data and the physical vulnerability of electronic storage media in terms of institutional (cultural) sound tradition. This is not just a technical question, it has an epistemological dimension as well. (26)

Coda: Material entropy of the signal versus symbolic (archival) endurance of sound recording

The video artist Bill Viola once pointed out "the current shift from analogue's sequential waves to digital's recombinant codes" in technology. (27) Sampling and quantizing of acoustic signals analytically transforms the time signal into the information of frequencies which is the condition for technical re-synthesis (Fourier transformation). Digitalization means a radical transformation in the ontology of the sound record - from the physical signal to a matrix (chart, list) of its numerical values. Media culture turns from phonocentrism to mathematics. All of a sudden, the archive recurs, but in a more rigid shape than previous scriptural record offices: the techno-mathematical archive. When the transfer techniques of audio carriers changes from technically extended writing such as analog phonography to calculation (digitization), this is not just another version of the materialities of tradition, but a conceptual change. From that moment on, material tradition is not just function of a linear time base any more (the speed of history), but a new, basically atemporal dimension. Against the noise of the physical world, techno-logical, that is “digital” culture poses a negentropic insistence, a negation of decay and passing-away. Digital copies of digital records can indeed be produced almost without loss of data (except the quantization noise). Music on Compact Disc or a digital video can be reproduced frequently with stable quality which was utopian in recent times of analogue recording on magnetic tape. The secret of this temporal invulnerability is that it is just (physical representations of) numbers which are written on the Compact Disc; even after a thousand copies thus a zero stays zero and one remains one. But past sound should not just be "restored" by applying digital filters; it rather wants to be remembered with all the traces of decay which has been part of its tradition, its media-temporal (entropic) characteristics must be archivized as well. Let us remain close to the physical record. If sonic articulation from the past is to be preserved digitally, its temporal (entropic) behaviour must be archivized as well - like the scratch, the noise of an ancient Edison phonographic cylinder which is achieved by over-sampling. Let us stay tuned to this non-archival sonicity.

Wolfgang Ernst

Wolfgang Ernst is Full Professor for Media Theories at the Institute for Musicology and Media Studies, Humboldt University, Berlin. He studied history, classics, and archaeology; in the 1980s and 1990s his research focused on theory of history, museology and the cultural archive before growing into the emerging discipline of media studies. His current fields of research embrace time-based and time-critical media, their techno-mathematical aspects and their "sonic" qualities. His monographical writings include: Medium Foucault. Weimarer Vorlesungen über Archive, Archäologie, Monumente und Medien (2000); Das Rumoren der Archive: Ordnung aus Unordnung (2002); Im Namen von Geschichte: Sammeln, Speichern, (Er)zählen (2003); Das Gesetz des Gedächtnisses. Medien und Archive am Ende (des 20. Jahrhunderts) (2007); Chronopoetik. Zeitweisen und Zeitgaben technischer Medien (2012); Gleichursprünglichkeit. Zeitwesen und Zeitgegebenheit technischer Medien (2012); Digital Memory and the Archive, edited and with foreword by Jussi Parikka (2013); Signale aus der Vergangenheit. Eine kleine Geschichtskritik (2013).

( 1 )

Hal Foster, 'An Archival Impulse', in: October 110, Fall 2004, 3-22

( 2 )

Quoted from the "call for papers" to the current Mnemoscape issue.

( 3 )

See W. E., Aura and Temporality: The insistence of the archive, in: Quaderns Portàtils (Portable Notebooks) no. 29, e-book, ed. by MACBA (Museu d'Art Contemporani de Barcelona) 2013;

( 4 )

See Bruno Lessard: The ANARCHIVE Project. In: Convergence. The International Journal of Research into New Media Technologies 15 (2009) no. 3, 315-331

( 5 )

See Eivind Røssaak (ed.), The Archive in Motion. New Conceptions of the Archive in Contemporary Thought and New Media Practices, Oslo (Novus) 2010

( 6 )

Friedrich A. Kittler, Gramophone - Film - Typewriter (transl. Geoffrey Winthrop-Young / Michael Wutz, Stanford, Cal. (Stanford UP) 1999: 117

( 7 )

Adrian Mackenzie, The Mortality of the Virtual. Real-time, Archive and Dead-time in Information Network, in: Convergence vol. 3, no. 2 (1997), 59-71 (61)

( 8 )

For a media-archaeological analysis of the World Wide Web see Alexander Galloway, Protocol. How Control Exits after Decentralization,Cambridge, Mass. / London (MIT) 2004. An example of anarchival order clustering which is the spatio-temporal configuration of techno-capitalism in current media culture is the microblogging platform for photo, text and video.

( 9 )

Some of these thoughts have been presented at the session "Digital Culture" within the Tate / RCA Collaborative Research Project Cultural Value and the Digital. Practice, Politcs and Theory, May 20, 2014, Tate Britain, London

( 10 )

Warren Weaver, Some recent contributions to the Mathematical Theory of Communication (*1949), in: Claude E. Shannon / same author, The Mathematical Theory of Communication, Urbana / Chicago (University of Illinois Press) 1963, 1-28 (7)

( 11 )

See Heinz von Foerster / Margaret Mead / Hans Lukas Teuber (eds.), Cybernetics. Circular causal and feedback mechanisms in biological and social systems. Transactions of the Ninth Conference March 20-21, 1952, New York, N. Y., New York (Macy) 1953, ”A note by the editors”, xiii

( 12 )

Some of the following arguments have been developed in my talk "The Sonic Time Machine. Explicit Sound and implicit Sinicity in Terms of Media-Epistemological Knowledge" at the Department of Aesthetics and Communication, University of Aarhus, May 22, 2014

( 13 )

Kittler 1999: 180

( 14 )

Kittler 1999: 15 f.

( 15 )

But caution, this is not an archive: As long as an algorithm is missing which rules the transition of sound provenience to permanent storage, it is just an idiosyncratic random collection.

( 16 )

See Mark M. Smith (ed.), Hearing History. A Reader, Athens (University of Georgia Press) 2004

( 17 )

See Bernhard Siegert, Das Leben zählt nicht. Natur- und Geisteswissenschaften bei Dilthey aus mediengschichtlicher Sicht, in: Claus Pias (ed.), Medien. Dreizehn Vorträge zur Medienkultur, Weimar 1999, 161-182 (175)

( 18 )

See Edmund Husserl, The Phenomenology of Internal Time Consciousness, trans. James Churchill, Bloomington, Ind. (Indiana University Press) 1964

( 19 )

Walter Ong, Orality and Literacy. The Technologizing of the Word, London 1982

( 20 )

Theodor W. Adorno, Zu einer Theorie der musikalischen Reproduktion, Frankfurt/M. (Suhrkamp) 2001, as quoted by G. Mazzola, Musical performance. Springer, Heidelberg 2010

( 21 )

Friedrich Kittler, Gramophone - Film - Typrewriter, Stanford (UP) 1999, 5

( 22 )

Karin Bijsterveld, Mechanical Sound. Technology, culture, and Public Problems of Noise in the Twentieth Century, Cambridge, Mass. / London (The MIT Press) 2008, 26

( 23 )

Michel Foucault, Message ou bruit? [*1966], in: same author, Dits et Écrits, vol. I, Paris (Gallimard) 1994, 557-560 (559)

( 24 )

( 25 )

See Jonathan Sterne, The Audible Past. Cultural Origins of Sound Reproduction, Durham / London (Duke University Press) 2003

( 26 )

See Arild Fetveit, "Medium-Specific Noise", in: Liv Hausken (ed.), Thinking Media Aesthetics. Media Studies, Film Studies and the Arts, Frankfurt/M. et al. (Peter Lang) 2013, 189-215

( 27 )

"Bill Viola,The Sound of One Line Scanning" in: Dan Lander / Micah Lexier (eds.), Sound by Artists, Toronto / Banff (Art Metropole & WalterPhillips Gallery), 1990, 39-54 (47)

bottom of page