Jussi Parikka interview, on ‘Digital Contagions’
MF: How do you figure ‘the body’ or the biopolitical in your discussion of viruses? Clearly it would be possible to simply fall into the trap of equating computer viruses with biological ones, to mistake the metaphor for the thing named. On the other hand it is possible to trace the ways in which the term has been used to mark a cross-over between categories that is about a kind of understanding of kinds of behaviours not delimited by material instantiation, for instance a certain kind of dynamic of proliferation,that makes the term meaningful. What are the stakes in following this through?
JP:Following a metaphorical line of thought from the beginning would have been the easy way out, writing an analysis of the metaphorics and representations of viruses in popular media. Indeed, that JP: was the way much of virus discourse was approached especially in the 1990s, analysing the translations and linguistic passages between diseases of bodies and diseases of networks. Naturally language has been an essential part of the creation of the so-called viral discourse, but I am keen on insisting at least on two things: 1) language and metaphorics should not be seen as primarily or solely signifying systems but as part of wider material assemblages and that 2) the biopolitics of computer systems is about many other things besides language as well (two related issues of course.)
MF: So firstly, following Deleuze and Guattari, language works as order-words, which is quite evident in the case of software. Whereas it would be interesting to approach software itself as an order-word (where the execution is a defining part of the event of computer systems), the linguistic acts that frame, stabilised and valorise software could be understood as such acts of power and knowledge that try to give a consistency to the contested questions of “what is proper software?”, “what is illegal software?”, “what kind of software and network events are allowed, by whom?” Here, as you note, it is also a question of cross-overs between categories, very tactical cross-overs indeed, of translating and smuggling elements from another, foreign realm to for example technological networks. Here “virality” can perhaps be used as a term that flags towards this virulence of trespassing categories, something I wanted to integrate intimately as part of the methodology of Digital Contagions.
Hence, the question of biopolitics of network bodies, the biopolitics of viruses and other software. I try to think this through the Deleuzian framework of allowing bodies to be of various kinds and scales: from bodies of humans, to bodies of software, networks, etc. Michel Foucault and people drawing from his work, like Jonathan Crary and Giorgio Agamben, have of course paved the way towards understanding the crucial mission of modern politics being not that of human being and their linguistic acts (their social life as rational, communicating beings) but as having to do with the “bare life”, the life beyond or in a way “before” human beings as metaphor-using communicators. The birth of modern media culture is one of tapping into the intensive animal reservoirs of the human being: for Foucault this referred to the biological features of the human being (as a species), for Crary, this referred to the new physiological experiments tapping into this human being as a fleshy, animal body. Rosi Braidotti has recently wanted to emphasize the animality of this layer by referring instead of “bios” to the concept of “zoe”.
What I wanted to do was to continue this line of thought to technological systems, and biopolitics of software, where the question was not reducible to what people say or think about software, networks, digital technologies, how the biopolitics of digital culture is not interested (only) in controlling human minds, but the intensive life of software, for example – taking the material assemblages as its object, in a way. Thus, this calls for an ethology of software, of looking at the objects and processes as affects capable of forging relations, making connections, interactions and exchanges.
MF: In writing about the cultural aspects of software there is a real imperative to technical accuracy. Firstly because if this is not achieved it makes the possibility of dialogue with those in the area primarily concerned with technical aspects quite difficult. Secondly, there is a kind of rigour required which is likely to produce new ideas rather than act as a blockage. How have you handled this in Digital Contagions, and how do you see this question developing?
JP: This is a question or an agenda that I learned to appreciate through German media theory, first via reading Friedrich Kittler, then Wolfgang Ernst among others. It also relates to what I just said about trying to think beyond the metaphorics of media culture and try to understand the more accurate expressions, techniques and ways of articulation that a medium might use beyond the human representations of it. So technical accuracy is a question of ontology (an often banned word in cultural studies) but as you suggest, it has the potential of acting as a vector beyond the confines of disciplinary boundaries. Now I do not consider myself expert concerning the technical characteristics of computer viruses, but related to the biopolitics question I see that a meticulous interest in this field is of crucial significance.
What recent years of approaches to networks, software and computer systems have achieved is a growing understanding of the questions of immanence of technology and power. Instead of bracketing the materiality of technology in the cultural studies agenda of ideology, much of the research done has succeeded in demonstrating how technologies in their very materiality channel and refashion power relations. They are not only second order results of any “social” struggles in the sense of “social” being something removed from the material. An understanding of the technologies at hand is a key prerequisite for an understanding of what kind of new modulations of reality we are dealing with. But I would not perhaps too swiftly call this as an aid in communication or dialogue, because it supposes that the concepts, or the “understanding of technologies at hand”, are transparently stable objects. Instead, also this material level is very much contested and what is crucial to me is not only an approach that takes into account of what kind of technologies we are dealing and tries to find the truth of e.g. software there but an approach which discusses this in terms of materiality that is continuously processual, not pinned down to a certain essence whether technological or social. Instead, we are continuously dealing with processes that are translational, in the process of being defined and across platforms. Not every computer scientist or anti-virus researcher is happy with what I write about viruses, quite the contrary, I’ve encountered arguments that I do not understand the technical reality of what I am talking about and that taking into account e.g. alternative voices in fiction is just leading my analyses astray. Again, in such statements we find the desire to pindown the truth of computer viruses to a certain technical knowledge, cut off from the translations and processes this weird overdetermined object is articulated in. So in addition to valorising technical accuracy, I would like to insist more widely on the materiality of the phenomenon at hand, a materiality that is irreducible to “agreed on” technical characteristics, a materiality that takes into account the various levels of relations and definitions of networks and software. Rigour is a good word, as it connotates a different thing as “technical accuracy”: it takes into account that one can be attuned to the materiality of the networks at hand, but without taking such a stance that “first you have to sort your facts out, then you can make your interpretations of those facts.” If we could do that, we would already have a fixed framework for those interpretations.
MF: Your period of study of computer viruses ends in 1995. Could you say something about why you choose this period as being significant, and what were the aspects of viruses you’d like to have covered in the subsequent period?
Yes, the period my study covers is approximately from the early computer era after the World War II onto approximately the emergence of the “popular Internet.” In a way this is of course stupid to stop there when the Internet was becoming an everyday reality instead of just a discursive promise of a networked future that was proposed in various platforms from professional computer journals to popular culture. But it is also because of this seeming paradox that the earlier period is interesting. For example the security discourse around viruses emerged at the end of the 1980s, and much of the techniques, tactics, and framings we use to make sense and control malware were not so evident at first. Focusing on the earlier period gives one access to the actual genealogical emergence of the phenomena and a truly historical take on the forces that gave consistency to the viral and other forms of malware. Here, one sees the recurring tropes emerging, like the curious insistence in computer security discourse to move from technical issues to social ones. So continuously, from 1960s on, you have the idea of “it’s the human being that is the problem, not the computer or the program” being articulated, similarly as the idea that “there is no good virus”, since the 1980s. Or then the continuous doom laden adverts and discourses warning of “data loss” at least since the early 1980s before viruses; “data loss disasters” to databases and personal computers due to various reasons from natural phenomena like the lightning to malicious intended crime, all of which in a way “paved the way” for viruses to fit into the already stated fear of data loss as a key danger of digital society.
Also, in terms of programs, much of the interesting stuff was done already in the 1950s and 1960s like the Darwin program or early rabbit batch jobs in mainframes. One of those, from 1966, included a RUNCOM command script repeating itself continuously which would then constipate the system (as David Ferbrache suggests in his ‘A Pathology of Computer Viruses’ book). Or how Kevin Driscoll attributed the emergence of viruses not to a specific program but to a short piece instruction, MOVE (Program Counter) –> Program Counter + 1, where the ”virus” is less a program entity than an instruction that is continously on the move to the next memory location. Besides being curious examples of an ”archaeology of the computer virus”, such processes should be taken as compelling issues that force us to think the digital culture in a historically tuned field.
This choice to focus on the pre-1995 period is in accordance with my belief that historical and temporal perspectives can bring forth novel rewirings and short-circuitings for present discussions and practices. Hence, Digital Contagions analyzes the media archaeology of this specific computer accident as a symptom of a more abstract cultural diagram. The digital virus is not solely an internal computer problem but a trace of cultural trends connected to consumer capitalism, digitality and networking as the central cultural platforms of late twentieth century as well as the media ecology and the so-called biological diagram of the computer where the biological sciences are actively interfaced with computer science often with a special emphasis on bottom-up emergence. Again, we are moving much beyond the more narrow take on recent years of “actual” viruses, and focusing on the archaeological transcrossings of the phenomena. Despite the often-stated idea that cultural studies is an approach that takes historical perspectives at its core, I would claim that a major part of this is done in a very vague fashion, neglecting concrete historical examples and work or reducing them to curiosities. Another way to consider historical perspectives is to contrast them with the affirmative perspective of becomings, which repeats a certain Deleuzian dualism: history as the regime of the State Archive and becomings as ahistorical creations. Instead of repeating this dualism, I wanted to approach the possibility of media archaeology as a nomadic cultural analysis, where “history” is not a marker of “already beens” but a potential, a potentiality that can be rewired into new assemblages of the future. Historically tuned cultural analysis cannot be reduced to a status of repeating the sources, but can be seen as one of summoning events as Foucault coined it.
Of course, this does not mean that focusing on recent years would not provided fresh perspectives. But there are people working already on this, like Tony Sampson from University of East London, finishing a book on cultural theory and viruses. I myself would have definitely refined my take in relation to e.g. botnets, wrote a few more words on net art viruses (which I am doing for the forthcoming Spam Book) and also more carefully would have covered the phenomena of terrorism.
MF: With viruses aimed at mobile phones running Symbian such as Cabir and Cardtrp, the latter which also crosses between Windows machines, the platforms for viruses are becoming more diverse. But with events such as the attacks on Estonian networks and the apparent existence of very large scale botnets, the broader category of ‘malware’ is itself becoming more infrastructural, more built into the internet. How does the figure of the virus work in this wider context?
JP: For sure, the notion of the “virus” or “viral” is in danger of becoming a floating signifier, a notion used for anything related to malware or in contrast, anything “cool” and “rebellious”. This relates to the earlier question concerning technical specificity which can be seen as one way of getting oneself out of the swamp of metaphoricity and vagueness and looking into how on the material level certain types of software function. My point was in general that malware has from early on been infrastructural to the Internet and network societies, this has been evident from early computer security texts since the 1960s on. The shift from protecting computers from human beings to protecting them from malicious software started around 1970s, and the notion of the incidental nature of the viral with networks feeds nicely into this as well. This is why I used the notion of the “universal viral machine” from Fred Cohen, the computer virus research pioneer: to underline that in the age of networked computers, viruses in Turing machines can be thought of as potentially semi-autonomous processes, a ‘“Universal Viral Machine” which can evolve any “computable number”.’ Cohen describes in his early work from 1980s (his PhD thesis came out in 1986) a weird world of computer processes without human interventions, there is not much mention of “intentions” or “social constructions” of computers, but anonymous processes, Turing machines, evolutionary sets and also e.g. “Universal Protection Machines” that are aimed to combat the Viral Machines by maintaining subject object matrixes, sequences to be interpreted, the rights of subjects to objects, scheduling of processes etc.
But we should not be blinded to think that because of the underlying Turing sequences, the processes are not system specific and material. Botnets are not the same thing as early 1990s viruses, nor is the 1988 Morris worm the same thing as current network worms that can spread across the globe in a matter of hours. Several of the early viruses got “extinct” because of technological obsolescence, their ways of proliferation via e.g. floppy disks becoming obsolescent. Much of the talk surrounding the new viruses suggests at least implicitly that viruses and their programmers are continuously finding new platforms and almost universal ways of propagation like via the Bluetooth in mobile phones. However, even though not being an expert on this issue, I understand that for example the Cabir worm relies much on the “kindness of the user” than on a system vulnerability, as the recipient has to accept to receive the particular piece of data package before the worm spreads. With Cardtrap, despite its malicious payload, it does not seem to work even with all Windows machines where the phone memory card might actually be carrying the Trojan but the autorun file did not, at least according to F-Secure information, function on Windows XP SP2 and Windows 2000. Again, much more than demonstrating the universality of the viral in the sense of cross platform spreading (which in a way is true as well) this also refers to the metastability of programs and their environments and how easily “things just don’t work” so to speak. This is the reason why Mark Ludwig flagged in the 1990s already that true evolution in software environments – at least the everyday environments like with Windows – is quite a far-fetched dream (or a fear) as the operating systems and software are just too unstable to allow for a random mutation that would work.
As for botnets, it’s the zombie side to them that is interesting. Eugene Thacker has been digging into the zombie world of contemporary biopolitics, looking at contagion and transmission through this figure of the undead, the life on the border of zoe and bios. Again, I would use the idea of the botnet to illustrate how power operates (also) on the level of ahuman technical, before or between the human social bind. Capturing computers in a zombie network is not reducible to a work of ideology, or as in the case of attacks against the sites of Estonian government and other public bodies to a work of international politics (even if it also was touched as the diplomatic relations between Russia and Estonia were involved), but a whole another layer of politics, working at the level of infections, software and networks. A lot of the analysis surrounding the attacks was seeing this from the viewpoint of international relations of two governmental bodies, but more interesting are the sub-governmental forces in action and also the sub-social forces that were harnessed as part of the international politics.
MF: One of the things that is interesting about viruses and other related kinds of software is their approach to computers and networks as a set of experimental zones. Towards the end of your book you mention Stefan Helmreich’s call for a ‘playful science’, showing how Artificial Life can correspond to this. At the same time, Viruses seem to have a slightly different form of playfulness to them. If we can adopt the language of probability for a moment, we could say that because Alife, generally (aside from interesting working done in evolutionary hardware, or in aspects of CrystalPunk work) tends to remain within well-defined boundaries, that of the model for instance. Whilst it has the capacity of offering a ‘theoretical’ playfulness, its is limited to a particular scale of activity. Viruses on the other hand offer a fully ‘experimental’ that is, more multi-dimensional, unpredictable way of inhabiting and shaping the networks. It sets in play are sets of conjunctions that are not simply within the domain of the software per se. The focus on malware tends rather to limit this. Your book calls for a more playful approach, where do you see the most useful historical resources for such playfulness? Which unexplored viral domains are most potentially interesting?
In a more straightforward vein, one could see my book as Foucauldian mapping of how the notion and powers of viral sets became territorialized and captured under the notion of malware, which acted not only as a repressive mechanism but produced a huge amount of books, advice, security instructions, manoeuvres, software etc. But to track this playfulness works a bit further on the issue. This actually relates to the question earlier you asked about why I stopped my analysis in 1995. It is just because the much more surprising stuff is found earlier, trying to follow the related strands of viral programming and the birth of network paradigms in computer labs. I was fascinated to hear from the early pioneers like Doug McIllroy, Vic Vyssotsky and Ken Thompson of their early experiences with computer ecologies of self-perpetuating programs. In a way, the obvious connection with early experiments had to do with the Cold War and security discourses, but I would say that much of the work done was not reducible to that functionality but also worked on another level of fascination with the expressions of these programs. For example, the simple game called Darwin that tried to out-populate the game ecology by “killing” other programs and spreading its own code is an interesting example. It was popularized later by A.K. Dewdney in Scientific American and now known as Core Wars. But what for example the above mentioned Ludwig flagged in his “black books of computer viruses” is that alife viruses are more or less dysfunctional. Due to the fundamental instability of most of computer systems, even small changes in code cause most likely only system crashes, no evolution. Hence, one has to deal with very limited scales, as you mention, and more interestingly speculate on the possibilities of for example evolving programs. It is a bit same thing as with artificial life art, where the genetically grown forms are indeed interesting and as an idea it has much to contribute, but besides the certain amount of forms “grown”, it starts to get repetitious (without a difference). Another problem in the whole artificial life virus discussion was the rigid way of dealing with the issue: to come up with a minimum qualifying definitions for an entity to be living (definitions adopted from observation of biological entities mostly) and then comparing this to computer viruses. Not a very interesting way to approach the issue – even though alife research has aspired to move away from this model-thinking onto a simulacra-approach, as Claus Emmeche suggested already some time ago. In any case, instead of merely following such paths, I wanted to propose a Spinozian ethological way of approaching “life” not as a substance, not as a form, but as an intensive life of affects, of interactions and relations where the life of technical bits is not to be removed from the life of other scales, or other assemblages. A small excursion: this interplay between ecologies of computational life and “social” life outside those systems is well reflected in the prize winners of the very recent VIDA 10.0 artificial life awards.
So life is not a metaphor adopted from biology and biology a model used to imitate the intensive code life of programs, but life becomes a movement, interaction and affects. This is the idea of playfulness as well: that the “ecologies” of media are not prefixed, stable natura naturata kind of mechanics in the service of capitalism, but also active virtual ecologies of natura naturans, of creation, probing and experimenting. To put it into Foucauldian vocabulary: let’s leave it to the police to decide whether the stuff really is alive.
As said, often the more interesting “living” experiments are the earlier, less researched experiments.
What also definitely would need much more research are the wonderful early computer ecologies of for example Nils Barricelli, Oliver Selfridge and Beatrice and Sydney Rome, all developing already in the 1950s systems that are relevant to the topic of experimental sciences of computational life. Even if not touching on viruses per se, they speculated on how to make ecological and evolutionary models work with a computational platform and how to make that kind of computation useful. Now if Cohen tried to figure out the usefulness of viral machines in the 1980s, these persons were speculating on this stuff already 30 years earlier! For example Barricelli did not want his work to be seen under the representational paradigm of computers modelled on life, but underlining that the stuff on symbiogenesis in computers is really there, as simulations. In other words, the simulation did not offer information on biological parasites and ecologies, but was an end in itself in offering a computer system that could work in terms of interdependencies, connectedness, symbiotic relations. As interesting are for example Oliver Selfridge’s Pandemonium experiments with semi-autonomous code of demons that “evolve” at least in a restricted way. Computation was understood there as a statistical mesh, a parallel processing based on the connected sum of “shrieks” every data demon of the system communicated to others. This also showed a system of distributed intelligence, as already Manuel DeLanda noted earlier, where such projects were seen as part of the genealogy of passing control from the human to distributed systems. In such a system, ideally, control “floats” from a demon to another which can take up on various functions, enter into flexible changing relations based on the global characteristics of the system that continuously feeds into the local relations of the demons. What is of course funny is how there is a curious correspondence between such computer system characteristics and the post-Fordist notions of e.g. work skills as branded by needed flexibility, adaptation to change, fluid communication…
Another theme is the experimental aesthetics of (technological) failure that characterise modernity. There is whole history of things breaking down, of course, and art has of course been one key practice of modernity where the failures of systems of technology, organisation and control have been catalyzed and experimented upon. This is the famous Paul Virilio’s notion of technical modernity: that accidents are incidental to their functioning. The accident of any system is a future horizon, a virtuality that might not ever actualize but it is still there in reality – often expressed only in statistics, worst-case scenarios and like, or then in simulated accidents by media artists. So much of the early avantgarde “media art” was based on exactly these impossible machines on the edge of breaking down, a Dadaist notion of technological modernity. One wonderful example would be George Perec’s 1960s radio play La Machine where a computer programmed to dissect and recompose in variations Goethe’s poem of The Wanderer’s Night Song. As Florian Cramer writes in his Words Made Flesh, Perec’s imaginary variation computer crashes and the input data turns into a program, working like an self-perpetuating email virus. I do not know whether I would agree with Cramer’s conclusion that this testifies of the superiority of semantics resisting syntactical programming, but I agree that this is an interesting experiment of aesthetics of failure, aesthetics of accidents. So perhaps the playfulness, in general, is trying to think beyond the most obvious choices, to think beyond the security discourse (which is a highly interesting topic of course) towards the experimental takes on viruses and accidents.
MF: Looking at art viruses, such as Biennale.py or those of Tomasso Tozzi in the 1980s there is clearly a further set of parallel imaginaries going on here. With tens of thousands of viruses in the wild, can you imagine or identify a particular strain working with a particular pattern of art methodologies?
JP: The art viruses, especially the Biennale.py project, fits nicely into this geneaology of aesthetics of accidents in its task to create an iconographics of malicious code. I think one of the fundamental successes of the project was to question the ontology of software and the distributed nature of the coded environment. On what level do micropolitics of software function, was an implicit key question of the project, which seemed to refuse a simple answer when distributing the code on t-shirts but also in expensive CD-ROMs etc. – while at the same time insisting on the harmless, invisible nature of the execution of the code. But beyond the way it was framed as part of art (as part of the Venice Biennale), what are the singular points to focus on?
I think Jaromil put it very poetically in the I Love You-exhibition catalogue when referring to digital viruses as a form of making (digital) language stutter in the manner Rimbaud and Verlaine made French stutter as part of an earlier challenge to transparent ways of seeing language. There is a threshold where code turns against itself and into a political gesture, or as Jaromil wrote: “In that chaos, viruses are spontaneous compositions which are like lyrical poems in causing imperfections in machines ”made to work” and in representing the rebellion of our digital serfs.”
From existing viruses in the wild, one could perhaps extract certain methodological principles. Much of them relate to finding the threshold just on the border of working and not-working: a virus that destroys completely the system is of relatively small use, instead much more interesting are the ones who are able to infiltrate the system and still keep it working (in a moderated form). That is, to find the threshold, the minimum level of a system before its flipping into a crash. In a way, this could be of course continued to the point of going over the threshold, of letting go of the control structures and seeing what comes up – of exposing oneself to the viral algorithms, as Joseph Nechvatal does with his viral paintings, which demonstrate how the viral noise is not antithetical to the ordered creations of art – virus itself can be turned into an emerging explorations of patterns in painting or in music. Here, variation becomes primacy, and the planned line and sounds are exposed to continuous slight variations of algorithmic kind. The methodological clue in general with viruses being: take any banal repetitious action without an inherent meaning, repeat the action or habit to the point when it starts to change, a point where the pure repetition produces difference from itself. This again can be seen as tracking the smallest differences and thresholds emerging in any systematic action and/or habit.
Another interesting theme is how the algorithmic logic of viruses feeds much beyond the computer code realm and takes advantage of the presumed sociability of human relations. Take the I Love You virus, a simple exercize in unfilled desire perhaps, feeding on the wish of getting a confirmation of love from someone. Or in another form, the gambler virus of early 1990s which forced the user play for the contents of the hard drive; answer incorrectly, and you will lose. This played with a certain mythology of a ”demon in the machine”, of the computer possessed which was a theme of Jodi’s early work of course (I think Alessandro Ludovico referred to their projects as insurrecting a certain alien presence in the computer which is a nice way to put it.) The virus examples mark the passing point or interfacing of the human being, but besides just focusing on the idea of the human being as the emotional, fallible creature, more interesting is to see the viruses, for example I Love You and other attachment viruses, as using to their advantage the habits of the user – of tapping into the presumed bodily habits where the meaning of an attachment is to open it etc.
Or then, to just track the parasitic movement and logic of the virus itself, as a way of exposing the dynamic logic of the net. Recently, the Google-Will-Eat-Itself took this parasitical logic of the Net to a new level by creating the paranoid-parasitical machine which draws money from Google to be used against itself. In a way perhaps this could be connected to the methodological ideal of “becoming imperceptible” and a move beyond identity politics. As argued by several Deleuzian writers, the becoming imperceptible of art is a much needed contrapunctual movement against the hegemony of representation analysis and identity thought where often only the only already recognized becomes an object of interest. How to come up with an action, experimentation that relies on the very notion of imperceptibility? An issue related to surveillance for sure, but perhaps also to art. In this context, Bertini’s Vi-Con is related to the notion of invisibility “Yazna and ++ are two viruses in love. They search for each other on the net, running through connected computers. Apart from other viruses, their passages won’t cause any damage to your computer […]. Theirs is a soft passage, invisible, and extremely fragile.” Software as soft, invisible passages, remaining imperceptible and again bordering on the status of existence – this focuses nicely on the question of what is the smallest meaningful act, what is the smallest perceived gesture that is taken into account.