ONE DAY I met someone I didn’t know I already knew. At the end of the first day of a communication studies conference in Albuquerque, New Mexico, a few years ago, I joined a group of attendees to go for dinner. We crammed into the back of a red Mitsubishi Precis hatchback and headed downtown. I started talking with the guy next to me in the car, Gilbert Rodman, who was doing a research project on the posthumous career of Elvis. I told him about my project on network cultures. After we’d shared our research war stories, we both had an uncanny sense that we’d met before, and recently.

And in fact we had met before. Just a few weeks earlier, my avatar “cdotc” had met his avatar “Brinkley” in a virtual environment for media researchers called MediaMOO. We had had a conversation not unlike the one we were having in the car, only this conversation had been in text only and we had been typing it on computers at opposite ends of the earth.

Obviously, this strange coincidence would have been impossible without the internet. As a technical network, it was the platform that supported our first encounter: a prime example of network society in action. But the internet was not the only network that helped bring about this uncanny coincidence. Each meeting was as much a product of the professional networks that we operated within as the technological networks that we used. Both the conference and the chat room were nodes, or attractors, in a global network of people working in the area of media research. We identified ourselves as media researchers and this shared affiliation pulled us towards the same spaces, virtual and physical. Other networks brought us together for the second time: a highway system, an aviation infrastructure and a circle of friends. Each of these has its own speeds and modes of operation but all served to smooth the paths that led to the back of that car in Albuquerque.

Perhaps the strangest thing about this event was that it wasn’t that strange at all. This kind of experience is becoming common in a world increasingly mediated by networks and correspondingly less by geographical location, hierarchies or longstanding contractual ties. The expansion of networks facilitates connections that are unbound by obligation, location or tradition. Network connections are not random in the sense of being based on chance. They promote random accessibility – links from any node to any other node. But this indeterminacy leaves networks open for flows and patterns to emerge based on common connections. Without our knowing it, such networks brought us together – twice.

THE GROWING POWER of networks has become, according to some, a defining characteristic of our time. We live in a network society and share a network culture. But networks are very difficult for any individual to perceive, since you only ever connect with a limited number of people (and things). You don’t know all the people those people know or all the people that they know, and so on. You don’t know all the vectors that brought breakfast cereal to your table, electricity to your toaster or email to your in-box. And yet your life is bound up in networks that trace invisible and complex connections for goods, ideas, energy, diseases and interpersonal connections.

So how are we to understand the nature and significance of networks? It seems a very abstract question. Even supported by many anecdotes and experiences like my own, the wider implications of these connections are elusive. Networks organise the world differently from the way hierarchies do: order emerges from multiple local negotiations rather than from designs imposed from above. Networks structure space in different ways from grids: connectivity is more important than proximity. The story of the network society involves both a new way of understanding the world and a way of ordering it.

The growth of information technologies is bound up in this story, but not in any simple way. In some ways, it is yet another instance where a new technology is adopted as a metaphor to explain everything. In the 17th century, the development of intricate clockwork technologies coincided with the Newtonian view that the universe worked with a similar regularity and order. Networks represent a very different image of a more complex and unpredictable universe. But computer networks are not just a metaphor. They facilitate, inform and accelerate other networks. Databases capture traces of networks in operation – in log files of transactions. Communication infrastructures allow real-time negotiations and ubiquitous connectivity. Computers embody techniques for analysing and visualising network patterns. These technological network infrastructures smooth and make visible the platforms on which other networks can operate. The reliability and speed of computer networks allow other actors to work with less centralised co-ordination and more responsiveness to changing conditions.

THE NETWORK THEORY that has emerged recently is itself a network of sorts, physically and theoretically distributed and diverse. The recent work that I survey in this article approaches networks from very different starting points: sociological, mathematical, historical and philosophical. Much of it is interdisciplinary – drawing from knowledge beyond its own immediate specialisation. And yet, in doing so, each piece betrays its own disciplinary history as it ventures, less convincingly, into more distant fields. Unlike social theory that splits into radical and conservative camps, left and right ideologies, and east and west territories, network theory exhibits a more complex politics in which agency is more dispersed and power, space and time are multiplied and made more complex. While each writer emphasises the growing influence of networks, each frames that change in a distinctive way: as historical global economic and social changes; as complex puzzles that can be explained mathematically; as cultural transformations in thought and material culture; and as increasingly complex relationships between human/technological hybrids. The combination of all these approaches does not give a single cohesive picture, but a series of often resonating, but untranslatable, impressions. Networks look different from every node.

One of the more substantial contributions to an understanding of the network society is the trilogy of books published by sociologist Manuel Castells in the late 1990s: The Rise of the Network Society, The Power of Identity and End of Millennium. He has updated each of these volumes already. Castells’ sociological approach spans large institutional changes on a global scale, with a large number of national and urban case studies.

Castells sees technology as quite literally revolutionary – manifesting in a global technical infrastructure the libertarian spirit associated with the milieus of California in the 1960s. While networks have been a longstanding structural feature of human societies, they have always lacked the co-ordination to achieve instrumental outcomes. Hierarchical command structures are more rigid but have been able to achieve goals more effectively. Once information networks are laid on top of other networks, it is possible to retain the flexibility and adaptability of networks, and also to achieve instrumental outcomes.

Network society is also partly a consequence of economic crises when the engine of growth for industrialism – improvements in production and distribution of energy – ran out of steam. The capitalist postwar boom went quiet and communist statism failed. Informationalism achieved growth by improving flows of information but also emerged with a dramatically changed global environment. Unlike the Cold War world, which was polarised between two geographically discrete ideological blocs, network society operates with infinitely complex structures and power relations. The war on terrorism is a struggle with the ambiguity of fighting across networks more than territories.

As well as presenting a large amount of empirical material, Castells develops theoretical concepts to make sense of the changes he describes. The examples he uses include global financial markets, the narcotics trade, technopoles (technological centres such as Silicon Valley) and megacities (headed by Tokyo, Sao Paulo and New York). Using such cases, he argues that the rise of networks is associated with a number of trends that transform collective experiences of time, space and power: “timeless time” and “the space of flows”.

Time is less regulated by the consistent rhythms of the clock, or even the body. Knowledge workers can work more flexible hours. People’s life cycles are more variable because they are living longer and managing fertility with birth control. Global corporations work across multiple time zones. Castells sees these as examples of the rise of timeless time. New technologies promote two distinctive senses of time: simultaneity and timelessness. Network connections link the globe to allow simultaneous connections across immense distances and time zones. At the same time, history loses its order, as hypertext and databases present texts from any point in time as a consistent multimedia universe, out of all sequence.

The concept of timeless time rings true to experience. Online environments such as MediaMOO facilitate conversations across time zones and also store biographical information about those with whom each user is interacting. But what was an academic experiment has since become popular culture. While MediaMOO has been practically abandoned, multiplayer PC games, such as EverQuest, internet real-time chat and networked games consoles have become very popular, creating a leisure culture of timeless time.

Another transformation Castells associates with the increasing significance of networks is that the “space of places” is subsumed by the “space of flows”. While Australia is not among his case studies, the patterns he describes are familiar. Places become defined less by their intrinsic qualities, or by their relationships with immediately surrounding regions, and more according to how they operate as nodes in global flows of information, goods and people. The fate of cities relates increasingly to where they sit in relation to these global flows. Among the consequences of these forces is that some areas and people are systematically bypassed. Innovations tend to emerge from well-connected and prosperous nodes and hubs.

The second book in the series concentrates on the growth in the power of identity, which Castells reads as a cultural reaction to networks. With declines in traditional sources of meaning, asserting one’s identity becomes a key way of making sense of the world. This change is apparent across the political spectrum, from feminist and anti-racism movements to religious fundamentalisms and nationalism.

Castells’s description of a global transformation into the network society is well argued and substantiated by historical evidence. But it lacks attention to the processes by which networks form and change. What are the cultural dynamics specific to networks? This question is possibly better addressed from quite a different region in the network of network researchers: experimental mathematics.

IN THAT CAR in Albuquerque, when Gil and I realised that we had in fact already met, we both exclaimed, “What a small world!”, as is customary in such situations. We weren’t aware at that time that there was a group of mathematicians who shared our surprise at such apparently random events, who were modelling this kind of event as a network phenomenon.

Among these mathematicians is Duncan Watts, whose recent book Six Degrees: The Science of a Connected Age tells the story of how this tradition of mathematical network theory emerged in the late 1990s. Unlike Castells, who begins with a detailed study of geopolitical changes associated with networks, Watts looks for network patterns he can describe mathematically. Where Castells starts by observing the world, Watts starts by imagining possible worlds and creating models of them. Information technologies are significant for both: for Castells they are a historical force for social change; for Watts they are the means of performing the complex calculations that model his imaginary networks and matching these models against data from actual networks. Networked computer databases also are a great source of such data.

Watts is particularly interested in using models to provide explanations for complex social events: from cultural/economic phenomena such as the dot-com bubble, to cultural fads such as Harry Potter, to the “small world” experiences that reveal unexpected social interconnectedness. How does an idea that necessarily begins in one location suddenly get taken up all over the world? What are the dynamics of collective opinion that allow markets to extend themselves beyond all rational bounds? What is it that makes one cultural product so successful through word of mouth alone? Are there really only six degrees of separation between everyone on the planet? In all these cases, Watts argues that the answer is not in the individuals but in the dynamic processes at play in networks.

One surprising outcome of these models is that it only takes a small number of long-distance links to model a world that operates as a “small world”. With computer simulations, Watts tries out simple models of worlds with different degrees of interconnectedness. In a world where people have connections with a limited number of other people, and all of these connections are local, there is no way that an idea in one location will spread to a disconnected location. But even a small number of links across to another location will mean that the whole system operates in interconnected ways.

As well as modelling whole networks, Watts models the nature of the connections between network nodes. A relatively simple example of the importance of the ways in which nodes connect is in epidemiological models tracing the spread of diseases. Different diseases spread through different vectors – air, contact, blood, etc. Each disease’s contagion profile affects how it spreads through a population. Even a small variation in contagiousness can mean the difference between an isolated outbreak and a pandemic. The critical threshold, when a local event suddenly becomes global, is often referred to as the “tipping point”.

Of course, ideas don’t spread in the same way as diseases, so such models don’t translate directly. So Watts draws on psychological experiments for principles about how people share ideas. One experiment shows that people tend to take up a new idea only when almost all the people around them accept this idea. Incorporating this principle into a model world produces a pattern of “information cascades”, where occasionally one idea is taken up right across a network. Most ideas die out early on. But sometimes an idea is initially percolated locally by a large enough critical mass of people to reach a tipping point, when it suddenly accelerates through a self-reinforcing spiral to be taken up globally. This model seems to fit the experienced dynamics of cultural fads and market bubbles.

Watts’s work, and other small-world theory, offers some compelling insights. It shows how very complex patterns can emerge from interactions between quite simple basic elements and rules. The information-cascade model also shows that cultural fads can’t simply be attributed to the exceptional genius of their creators. Rather, their success is to a large extent attributable to network dynamics. Similarly, the “power law” model shows that small differences in initial positions can make huge differences at a later stage – a network mechanism whereby the rich always get richer.

However, mathematical modelling of networks always involves reductionism. Most actual-world problems are just too complex to describe, so models are massively oversimplified simulations, lacking any historical specificity, even when Watts tries to incorporate sociological insights. They always beg the question of what it is that they are actually modelling, and how such information can be translated into the messy business of the social world. And a model of any single network doesn’t easily take into account its intersections with the many other networks that are always at play: there are no boundaries between social, biological, technological phenomena, and all affect others at all times.

Small-world models have little capacity to address the qualitative differences in the cultural dimensions of information cascades or any other phenomena. What is it about Harry Potter that resonates with our times? Universalising models lack any historical frameworks that would explain more subtle cultural shifts that have accompanied the growth of networks.

WRITERS IN THE humanities are better than mathematicians at analysing the texture of cultural changes associated with the rise of network culture. For example, Mark C. Taylor, in The Moment of Complexity: Emerging Network Culture, surveys the work of theorists of complexity working in the same vein as Watts, but talks with more philosophical and historical grounding about the implications of such network models. He traces the emergence of a new network culture, as it is manifest in changes through time in cultural artefacts in architecture, painting and literature as well as maths, science and technology.

He shows how buildings change from the rectilinear forms of the modernist skyscrapers to more dynamic forms like Frank Gehry’s Guggenheim Museum in Bilbao, Spain, (or we might add Melbourne’s Federation Square). Theoretical paradigms change from the structuralist emphasis on systematic, formal systems to post-structuralist emphasis on complex and dynamic processes that generate structures that are never stable or permanent. In the self-portraits of Chuck Close, Taylor traces a development from the hyper-realism of his early work towards the hyper-complexity of his later paintings. These later images are composed with fragmented grids that appear totally chaotic up close. These fragments only merge to form a cohesive image when viewed from some distance. Each of these cases is characterised by complexity rather than formalism, emergence rather than instrumental design. In these works of art, Taylor sees evidence of changing values in the wider culture.

More unsettling is the suggestion that thought itself changes dramatically with the growth of networks. Thought is not something that you simply do on your own. You don’t invent the language you speak – you borrow it from your community. Your thoughts themselves are composed on the basis of shared perceptions, concepts, techniques and frameworks. If the way that the materials of thought are distributed changes, then thought itself changes. Taylor argues that recent cultural and technological shifts have passed through a tipping point beyond which turbulence has become a constant: “As networks relentlessly expand, the mix of worlds, words, sounds, images and ideas becomes much more dense and diverse. When this media-mix approaches the boiling point, multiple cognitive and cultural changes become inevitable.”

Unfortunately, Taylor’s journey is more interesting than the destination. After talking so much about embracing complexity, his conclusion that business and education should join forces, and that those that don’t embrace this change will wither away, is laughably simplistic and self-serving.

Taylor’s conclusion that we are encountering an epochal change is echoed in much recent work on technology; and, of course, belongs in a long tradition promising a sudden ascension to a new order. Damien Broderick’s The Spike and Ray Kurzweil’s The Age of Spiritual Machines make even more extravagant claims about an approaching age when technological innovations will make the world unrecognisable. Pierre Lévy presents a more knowing utopianism in Collective Intelligence, in which he proposes that networks can support a fluid ecological democracy. Lévy’s project is to provide a vision that might influence future directions in politics and technology.

The inverse image of the utopian network theorists are the apocalyptic critics including Paul Virilio and Jean Baudrillard. Virilio fears the loss of “real time” in the wake of absolute speed. As technology escapes actual scale, all possibility of human control disappears. Baudrillard’s anxieties centre on the pre-eminence of simulations over reality. While both the utopian and dystopian approaches present strong images of the present and the future, neither squares up with most people’s actual experience of networks.

The changes associated with new media are subtler than these expressions of cultural anxiety or hope suggest. Some of the more convincing work on network cultures emphasises that networks are not some universal force, but are historically specific cultural forms. Robert Burnett and P. David Marshall’s Web Theory is an accessible introductory text that positions the World Wide Web within a history of media technologies. It emphasises how difficult it is to define the web as a single medium: it is a “loose web”, with sites ranging from personal home pages to corporate websites. This looseness crosses the conventional boundaries between private and public space, broadcasting and conversation, and between authors and readers.

Some other recent work reflects on how technology changes everyday experience, including that strange experience of meeting Gilbert twice. This was possible because of the quality of difference between meeting face-to-face and meeting online. John Potts and Andrew Murphie develop a philosophically inflected reading of cultural changes associated with technology. They argue that artists like Stelarc, who connected nerve stimulators to his body to allow others to move his body over the internet, play with the conventional boundaries between humans and technologies. These changes are not a total transformation of everything but a complex mix of large and subtle shifts that challenge us to renegotiate how we live.

BUT IT IS possible to explore the growing power of networks without making the universalising claims such as Taylor’s? Actor-network theory (ANT), which emerged from social studies of science in the 1990s, operates with a very broad conception of networks: long and complex linkages between human and technological actors. In this approach, though, every network is different and the connections between networks are as important as networks themselves. While some of the work in this tradition can be dense reading, its combination of close case studies and sophisticated theoretical analysis is compelling.

For a long time, students of technology and society were stuck in a conflict over the relative importance of social or technological processes. Does technology lead to social change or do social forces create technology? ANT is one way of resolving this impasse. It suspends the question of whether the cause of any event is an object or a person and subsumes both under the term “actor” or “actant”. Then it defines actants not by their intrinsic attributes, but by their relationships with other actors (their position in a network). Rather than mapping networks of one kind (social relationships, computer networks, electrical networks, ecological networks), ANT investigates the connections that cross over these conventional networks to form actor networks. So, for example, online grocery shopping involves a network of alliances that bring together not only customers and retailers, but also farmers, fertiliser supplies, international trade agreements, and even non-human actors like plants, land and the weather. Changes in any one of these networks ripple through to all the others.

Contemporary experience is conditioned by enormously complex hybrids that chain together human, technological and natural actants. Information systems may be a growing part of these, but to ignore their connection with other actants is to stop the network analysis too soon. According to Bruno Latour in We Have Never Been Modern, though, claims about the special status of modernity were based on denying such hybrids. By creating a split between nature and culture, the “modern constitution” separated the world of men from the world of nature – objective scientific facts were seen as independent of the subjective realms of politics. The complexity of networks is reduced by making sub-networks operate as black boxes, in which the decisions implicit in a system’s design are set aside. ANT uses ethnography and cultural theory to break open such black boxes and to reconnect politics, facts, devices and the networks they compose. Latour’s study of a failed project to create a light-rail network in France in Aramis and Steve Woolgar’s ethnographic research into the development of a personal computer reveal that technologies have never proceeded according to plan.

While each of the approaches I’ve outlined offers a reading of the complexity of networks, there seems little prospect of unifying these into any single explanation. All of the examples show how much networks complicate the conventional politics of representation and accountability. The dispersal of power through networks displaces and defers any attribution of final responsibility that is at least theoretically possible with conventional legal and political institutions. Castells’s sociology documents recent dramatic global changes in space, time and power. Watts diagrams the mathematical dynamics of networks. Taylor and Murphie and Potts show how networks have implications for thought itself, which are manifest in the material culture of architecture and the arts. ANT traces the very particular connections between power, technology and knowledge that are always at play.

AS I FOLLOWED Gilbert and the other international communications researchers out of the Japanese car into a Mexican restaurant in Albuquerque, I thought little more about the forces that brought us all together. The growing network forces conditioning even the most mundane day-to-day experiences are almost as hard to perceive or understand as they are to escape. Knowledge about networks is as dispersed as the networks themselves. There is no way to resolve a network to a single centred point, nor any way to exhaust the chains of dependencies that link networks to each other. On the other hand, with no stable centres of power, a new force for change might emerge, as if at random, from any point.

References

Baudrillard, Jean (1994) Simulacra and Simulation, Ann Arbor: University of Michigan Press.

Broderick, Damien (1997) The Spike: How Our Lives are being Transformed by Rapidly Advancing Technologies, Kew, Victoria: Reed.

Burnett, Robert and Marshall, P. David (2003) Web Theory. An Introduction, London and New York: Routledge.

Castells, Manuel (2000 [1996]) The Rise of the Network Society, Massachusetts: Blackwell.

Castells, Manuel (2003 [1997]) The Power of Identity: The Information Age – Economy, Society and Culture, Massachusetts: Blackwell.

Castells, Manuel (2000 [1998]) End of Millennium, Massachusetts: Blackwell.

Castells, Manuel (2001) The Internet Galaxy, Oxford: Oxford University Press.

Kurzweil, Ray (2000) The Age of Spiritual Machines: When Computers Exceed Human Intelligence, New York: Penguin USA.

Latour, Bruno (1993) We Have Never Been Modern, Cambridge, MA: Harvard University Press.

Lévy, Pierre (1997) Collective Intelligence, New York: Plenum Press.

Lévy, Pierre (1998) Becoming Virtual, New York: Plenum Press.

Rodman, Gilbert B. (1996) Elvis after Elvis: The Posthumous Career of a Living Legend, London: Routledge.

Taylor, Mark C. (2001) The Moment of Complexity: Emerging Network Culture, Chicago: University of Chicago Press.

Virilio, Paul, translated by Julie Rose (1997) Open Sky, London and New York: Verso.

Watts, Duncan (2003) Six Degrees: The Science of a Connected Age, New York and London: W.W. Norton.

Woolgar, Steve and Grint, Keith (1997) The Machine at Work: Technology, Work and Organization, Cambridge: Polity.