Dynamic Media Network

(draft schematisation for New Network Theory conference, 28.06.07)

This paper emerges from a background project that I have been unsystematically pursuing for the last 3 years or so. Various bits of it appear throughout texts – ‘Theses on Distributed Aesthetics: Or What a Network is Not’ (with Geert Lovink) and a more recent piece ‘Welcome to Google Earth’. In these essays I realise that I have been trying to understand the interplay of two aesthetic forces or vectors in network cultures – the pole of customisation, homogenisation and atomisation and the pole of collective enunciation, production and distribution. Not that these are ever poles apart in contemporary network cultures.

For a while I have thought about this as a project concerned with ‘distributed aesthetics’ but I have more recently begun to conceptualise it as ‘an aesthesia of networks’. This working title gathers into it the ideas of Castells, Terranova and Rossiter who have all argued that networks are constituted in the very tensions between the singular and collective, net and self and intensive and extensive processes and flows. Hence there can be no coherent, global ‘aesthetics of the network’. And yet there are collective and shared experiences – aesthesias – of networks. The most common experience of contemporary networks perhaps being repeated cycling through euphoria and boredom.

There are also recurring patterns that regulate the aesthesias of networks such that their hetereogeneity or singularitiy ends up being siphoned into a neater ‘package’ of network functionality. One of these operates by packaging the network as image and takes the form of the vectoral diagram of networked connectivity. This has come to function as a dominant image of and for networks.

who owns the internet? by Ben Worthen, Bill Cheswick

Lufthansa IT infrastructure

The repetitive and ubiquitous circulation of these kinds of diagrams of connectivity is striking in itself. But it is the aesthetic implications of these in which I am most interested. For I want to suggest that this diagram’s status as a kind of meta-image of networking is literally anaesthetic – numbing and disengaging from the chaotic and experiential engagements in networks. The node-link schematic lulls us into a kind of comotose state about the socio-aesthetic-technical assemblages that ennervate network cultures. What I want to suggest is that the far-reaching distribution of this image of distributed networking operates as a homogenising force that attempts to erase disjunction, relationality and temporality from our perceptions of/in networks.

Luckily, however, network visuality is not such a flatline! There are many examples of how individuals, online groups and environments are providing different approaches to the image in the network. I want to provide some examples of these later in this talk and to revisit the nature of these alternative images. Rather than trying to classify these images through a visual taxonomy, I will instead focus upon their divergent nature. In so doing, I want to invoke Walter Benjamin’s analysis of allegory in The Origin of German Tragic Drama. For Benjamin, allegory was not so much something to be found contained within a particular text or image and systematically interpreted. Rather his approach to baroque allegory was to understand it as a mode of seeing or reading predominant throughout the European seventeenth century but also potentially resonating with later historical/cultural conjunctions. Baroque allegory inhabited the sphere of everyday visuality – the domestic, the familiar, the street scene – and unfolded via contingent associations between its metaphorical elements, often moving from one element to another in unexpected ways. He compared this twisting variability of baroque allegory with the function of the symbol in art and literature. The symbol’s function was to preserve representational homogeneity – to always mean the same eternally.

I wonder whether this might not be a useful comparison to import into what I have to say about the ways in which the diagrammatic (rather than Benjamin’s symbolic) and the allegorical differ in network visuality. I think this may be a useful way to think about both the role of network diagrams and the role of alternative imagings of networks that I want to unfold today. These latter imaginings evoke a mode of visuality operating via divergent, disparate, everyday and surprising associative pathways. I think we find this allegorical mode in direct images of the Internet and its cultures, for example:

An allegorical map of online communities

but also in the attempts to stretch the diagrammatic mode to unfold the shifts of connection and disconnection that comprise the political dimension of networks. I am thinking here of the work of the artist Mark Lombardi who famously portrayed the money that filtered from the Bush family oil investments in the US into the Middle East and eventually was redistributed to the Bin Laden familiy’s attempts to rearm and refinance sectors of Iraqi society for their own interests:

george w. bush, harken energy, and jackson stevens c.1979-90, 5th version, (detail)

More recent examples of a ‘stretch’ of the diagrammatic mode come through visualisation software such as Digg Swarm, which dynamically updates the clustering of users’ ‘interest’ in stories posted on the Digg social aggregation news site. I think what we have here is a kind of becoming-allegorical of the diagrammatic. Of course it’s also the case that the incorporation of both clustering and tag clouds as attempts to make the diagrammatic more expressive in Web 2.0 design re-asserts a kind of visual homogenisation where the ‘clustered’ and/or buffed-up tag comes to visually dominate and other variables in the image plane easily fade…

I’m also especially interested in a kind of emerging web visuality that develops through a mash-up of the diagrammatic and the allegorical by layering geodata and imaging in conjunction with personal and collective data and imaging:

where’s george? mash-up

It should be clear then that I am using a conception of the allegorical here that broadens Benjamin’s to examine engagements with image-making in network cultures that have an everyday (sometimes even banal), contingent and divergent nature to them. I am aware this may prove to be too broad but I think its better and, in fact, crucial to cast the net wider in the present moment given the kind of grip the purist articulations of the network diagram has on contemporary networked visuality.

What, then, do I see as the problems of the diagrammatic mode for the visual cultures of networks? And why, subsequently, do I think we need to reinscribe the importance of the work of allegory in the age of informatic supra-production? It is not so much that the image of diagrammatic connectivity represents networks in bad or good ways. Rather, I want to suggest that this form of diagram has come to function as a network meta-model, laying out the conditions of possibility for the experience, the aesthesia of networks. Its limits are those that C.S. Peirce noted about the diagram as a form of mathematical notation – that it says nothing about disjunctive information, existential statements (that is the conditions that are fundamental to its operation as a notational system), probability or relationality. In addition Mat (Wal-Smith) has pointed throughout this blog to a number of issues concerning the planar-linear-spatial problems of contemporary network visualisations. Namely that these occlude the folded histories of actual interaction in/of the network. As he suggests in his post on Jess Kriss’ History of Sampling visualisation, the visualisation channels our mode of interacting with the historical data inputted about sampling. The visualisation draws planar graphs of the use of a sample in a piece of music but not how a sample might act as a catalyst for our relationships with the histories of music or to further processes of musical sampling. Hence we end up not with a history of the processes that are sampling but rather a history of samples (bits of trackable data).

What I want to do is think about this kind of processual semiotics endemic to contemporary media work, especially electronic music, as a mode of understanding network imaging. Another way to put this would be to pose the question of how images in networks are constitutive factors in network processes, flows and their regulation. First, I want to look at the domination of the diagrammatic image of distributed communications first sketched out in Paul Baran’s 1964 RAND memo (image to come). The circulation and repetion of this kind of diagram as a network map, mnemonic and actualisation now dominates the visual landscape of networking, informing social network analysis, network visualisation and net aesthetics. And then second, I’ll look at the ways in which the diagrammatic gets redrawn and mashed via allegorical network visuality.

When I talk about the processual semiotics of networks I mean to invoke not so much the tradition of interpretative semiotics that we may be familiar with via Sassure, Barthes and psychoanalytic theory. Rather I want to understand the diagrammatic via, as I have already mentioned, Bertrand Russell and Pierce and the ideas of processual semiosis that appear in the work of Felix Guattari.

I’d like to proceed by looking at Baran’s diagram in the context of his memo to RAND. I then want to make some general comments about how these kind of diagrams function to manage and organise our perception and engagement with networks in the contemporary moment – ie as a way of regulating network aesthesias as ‘an aesthesia’

The mythology associated with this diagram is that it represents the genesis of the digital network as sustainable in the face of nuclear attack. As the story that accompanies ‘the origin of the Internet’ goes: it was this distributed diagram allowing and attack on one node without meaning the whole network would come down. This diagram is often historically associated with the early 4 node hook up that initialised ARPANET in ’68/’69 and in fact the period and research culture overlaps certainly justifies the association:

Paul Baran’s diagrams of communications systems

Hence the distributed communications system somehow acts as the ‘blueprint’ for the emerging connectivity of academc and military networking in Cold War USA.

However, in an interview between Baran and Stewart Brand in 2001, Baran himself comments on this myth of Internet origins, insisting that it was not the connectivity of network nodes as demonstrated in the distributed communications diagrams that was at stake in sustaining resilience to nuclear attack but rather the flow of information and data via packet switching that would be essential for deciding both sustainability and strikeback capabilities for the network. (See the interview in wired)

This is an important distinction because it indicates that Baran was not so much invested in the realisation of this diagram as a blueprint for the network but rather was focused upon network processes – the capacity of data to divide up, rearrange and reassemble itself as it moved around connections – in other words, packet-switching. There is some authorial revisionism going on here. If we look at Baran’s original 1964 memo, he clearly states 2 criteria for post-attack survivability: both the percentage of ‘stations’ (as he calls them) left after attack and their ‘electrical connectivity’. But perhaps what Baran has in mind in the later interview ‘revision’ is that networkability – what he calls ‘the synthesis of a communication network’ as distributed (and what I am understanding as the technical and social capacity of distributed communications to be constituitive elements in network formation) – is not reducible to the actual physical infrastructure that ‘joins’ the dots in a network.

As has been repeatedly the case in the history of the implementation of information theory – especially in the history of its military applications but also in its migration into other disciplines such as media and communications studies – nodes, senders and receivers have been hypostasised to the detriment of investigating the processual movements of data and peoples. As it turns out, we have to understand Baran’s diagram and memo through both the poles of the hypostatic and processual. On the one hand, he is clearly interested in accounting for the precise ‘level of redundancy’, as he calls it, required in a network for it to function after severe physical attack on actual communications stations. This necessitates pushing the diagram through a series of graphs to calculate what number and level of nodes are needed initially for it to survive a severe attack on its nodes. On the other hand, after a certain amount of reduplication or redundancy of nodes the distributed network survives even a heavy loss of its actual infrastructure because of its array formation:

Baran’s diagram for array formation – a ‘process’ diagram

Baran is, then, equally interested in how the processes of distribution continue in the post-attack scenario. For him, these processes are only possible if the network has already reached a level of production of redundancy allowing the duplicative array formation. And for him, the array formation is simply the precondition for maximum switching of packets of information to occur. The distributed diagram, then, is not a blueprint for how to build a network – although there’s no denying Baran was working to a military brief. Rather it is a set of vectoral preconditions necessary for the process of switching to occur; a process that is, for Baran, sustainable not only in the event of attack but also in the face of everyday network failures: ‘noise’, unreliable links, degradation and overload. It is little wonder that process is constantly overlooked in the visual depiction of networks as diagrams of connectivity. Again and again in Baran’s memos network processes are entwined with a kind of implicit understanding of the aesthesia of networked inefficiency and breakdown. These problems of defective connections and systemic failure are hardly a vision of imperial preparedness for the nuclear age!!

At least part of the problem with the overlooking of the processual in network visuality lies with how we understand the representational status of diagrams and the historico-discursive forces shaping that understanding. In particular, I am thinking of the legacy that diagrams inherit from mathematics and syllogistic logic. Both Euler and Venn diagrams were developed to visually demonstrate syllogistic logic (example). However, as the analytic philosopher Bertrand Russell pointed out in 1923, there is a ‘vagueness’ to the diagram which in endemic to the problem of representation (Russell, 1923). Rather than the diagram simply acting as a one-to-one form of representation (as other forms of representation in mathematics such as algebra might), its spatiality frequently means that it acts in one-to-many mode. Hence for Russell, its ‘vagueness’ or rather its potential to be representative of the multiple and the variable. So, for example, this vagueness means that the spatial relations between objects in a diagram can be used to represent relations between objects in some other domain. Baran’s distributed communications diagram could be a diagram of ARPANET connectivity but it could also be a diagram of Lufthansa IT networking.

The diagram is therefore not a set of instructions – a blueprint – for mapping or building relations between objects. It is instead a representational mode that hooks one class of objects – perhaps links and nodes – to another class, potentially peoples, cultures and their processual relations within networks. This, of course, is why the network diagram is so thrilling – its spatiality and vagueness harnesses the potential to make it work as a representation of something it is not. The problem is that while the potential to transpose from map to ‘territory’ is one of the diagram’s visual attractions, we would do well to remember that this transposition is only a product of representational vagueness rather than accurate correspondence. In other words, if we really believe that the network diagram provides us with an accurate depiction of networks, then we are forgetting the very relationality of both diagram and network.

It’s also important to remember that the history of diagrams within the 20thc development of logic is a contested one. In particular, the interventions of Peirce into diagrams as a mode of logical reasoning can be seen as both a contestation of their representational limits and an attempt to enhance their expressive capacities. He extended the classic Venn diagram

by introducing new symbolic notation that could account for the presence of disjunctive information within a set:

This diagram allows for either the syllogistic proposition ‘All A are B’ or the disjunctive information ‘some A is B’ to hold in the one representational space

I don’t have time here to do Peirce’s extensions which also included attempting to extend the diagram to deal with logical existential statements…in fact for contemporary logicians Peirce’s extensions ended up becoming too visually complex and, since the 1990s, work on the diagrammatic mode in logic has had a strong focus on returning to visual simplicity. That’s perhaps unsurprising in the context of the broader visual culture, which I have also been attempting to chart in this talk, and which is underwritten by the seduction of the clean diagram as meta-model.

But what I am also interested in is the possibility that the diagrammatic mode can be deformed and shaken by the processual – and here I mean two kinds of deformation that are never far apart from each other in network cultures. The first I’ll call a kind of intensive deformation, which is catalysed somewhat by the Peircean project but is taken up again in the work of Guattari. Here the diagram tries to unfold its vagueness or what we might also call its virtualities – its potential to become other, its potential to move to other rhythms. In this kind of deformation of the diagrammatic mode what is at stake is the diagram as dynamic, the diagram as process.

A diagram by Brian Holmes that attempts to work with the processual relations involved in the shaping of new subjectivities of collective enunciation

I think this attempt to stretch the diagrammatic in processual ways is a strong direction as network visualisation attempts to come to terms with the intensive dynamicism of Web 2.0. It’s what we see happening in the Digg Swarm visualisation. It’s also what we see happening as node/link diagrams are subjected to weighted/dynamic mapping tools (thanks again Mat for pointing me toward the Middle East Power maps and toward Fidgt).

A snapshot of the Fidgt visualiser, which works by aggregating tags from users’ web accounts such as Flickr and lastFM. Entering your account into the Fidgt visualiser then aggregates other users with the same tags into your map of ‘use’ visualisation once you deploy a tool called a ‘Tag magnet’

I don’t necessarily want to endorse the ‘social networking’ claims here nor the questionable ethics of popularity/wisdom of the crowd behaviourist psychology that accompany the topology of tagging and weighting in Web 2.0. But what I do find interesting about what has happened to the diagrammatic here is that there is a notable shift from diagram as notation and representation (with all its attendent problems of spatialisation and location) to diagram as activity and process. What kind of an aesthesia does this embody and generate? A networked aesthesia of plasticity – potentially collaborative, generative of new problems for thinking and engagement but also collapsing, deteriorating under the weight of the endless generation of its own redundancies.

Finally I want to think again of another possibility for network visuality, which I touched on briefly when referring to the idea of web mash-ups of the diagrammatic and the allegorical. In the where’s george? mash up I showed previously, the mash is produced by overlaying the connective diagram with Google Maps. And this is of course where much of the mashing in networked visuality currently occurs – using Google’s API capabilities to embed its maps into user-generated data. Here we have a mash-up of locative data with data flow…and in some ways this is reminiscent of earlier web projects (many of which are archived in the Atlas of Cyberspace site) that attempt to provide a geospatialisation of network generated exchange and interaction.

But these could also be understood as a mash between the everyday and associative relations produced or generated by the collective exchange of peoples in networks, on the one hand, and the vectoral packaging of relationality into the data template on the other. It is in this sense, that I speak about a mash-up of the diagram and the allegory in network visuality (recalling Benjamin’s comments about the incipient wandering and everydayness of the allegorical as well as his ideas about synthesis as the ongoing presence of tensions and of the baroque as amode which comprised extremes in aesthetics). What I think we need to do is work at the potential for both the disjunctive (diagrammatic expanded in the direction of its expressive capacities) and the temporal (allegorical as a mode of unfolding historicity, everyday network realities) to play a more overt and generative role in our images and imaginings in networks. This may help us to actually produce networks that are less templates for relations and more ongoing projects that explore new relational forms for social collectivities in network cultures.