Internet: Towards a Holistic Ontology       Contents | Introduction | Chapters 2 3 | Conclusion | Notes | References | Appendices A B

Chapter 1

This chapter will investigate the degree of analogy between the Internet and the concept of the rhizome, a "style of thought" (Goodchild, 1996: 85) introduced by Deleuze and Guattari in the opening passage of A Thousand Plateaus (1987).1  To oversimplify, the rhizome functions through lateral proliferation rather than tree-like vertical extension.  It was offered as an alternative to the constricting tree structure and dualism dominant in what Deleuze and Guattari call "state philosophy" (Massumi, 1987a: xi):

Closed upon itself, and holding the lid down on its own discipline, philosophy has often mistaken vacuity and ineptness for wisdom and rigor and the foul odors of inbreeding for signs of intellectual and moral integrity.  To open the lid, so that a gust of fresh air may come in from the outside, is not to waste time deciphering the signs of an upcoming end of philosophy.  (Boundas, 1993: 1)

     It is important not to see the rhizome in binary opposition to the tree, as Boundas seems to suggest by the title of his chapter "Rhizome Versus Trees" (1993: 27).2  Deleuze and Guattari themselves caution that "the root-tree and canal-rhizome are not two opposed models" (20).3  The concept of the rhizome was set up precisely to challenge dichotomous branching; yet by placing rhizome against tree, how can we ignore the self-contradictory dualism?  How do Deleuze and Guattari propose to get their rocket — or line of flight, as it were — off the ground?  They call forth

[t]he problem of writing: in order to designate something exactly, anexact [sic] expressions are utterly unavoidable.  ...  We invoke one dualism only in order to challenge another.  We employ a dualism of models only in order to arrive at a process that challenges all models.  (20)

This is more than an instance of rhetorical "stuttering" through which Deleuze is known to arrive at fresh philosophical concepts (Boundas, 1993: 1); for due to the "multiple entryways" of the rhizome, it is possible to enter the rhizome through a tree-dualism, "assuming the necessary precautions are taken", including, paradoxically, an avoidance of "Manichaean dualism" (14).  The relation of the rhizome to the tree is not one of dualism but of territorialization, but we shall return to this following a description of the six principles of the rhizome itself.

What is a rhizome, and how is the Net a rhizome?

The six principles of the rhizome as listed by Deleuze and Guattari are as follows:

• connection
• heterogeneity
• multiplicity
• asignifying rupture
• cartography
• decalcomania

Goodchild reminds us that "these principles do not remain separate but mix with each other and proliferate" (1996: 85).  In other words, we should bear in mind that they do not function in isolation, but only together as a rhizome.  The rhizomatic method would ideally involve negotiating through all of these principles simultaneously so that they can proliferate in the reader's mind; but for reasons mentioned in the Introduction, the paper version of this thesis will stay with the linear mode of explication, while the HTML version will enable a nonlinear negotiation by the reader.  Rather than detail the principles and then say how they exist in the case of the Net, I will map each principle to the Net as I go along, in the hope of encouraging a form of proliferative mapping.

Connection and heterogeneity

Whereas the tree structure functions through a hierarchy and assigns ordering positions to its subjects (thereby constraining their relationship with others), in a rhizomatic system each point can and must have connections to all others, unconstrained by any bifurcating order (7).  The principle of dichotomy is abandoned, for it is self-contained and does not open to connections with other domains.  For example, Deleuze and Guattari criticize Chomsky's dichotomous linguistic model for failing to connect "a language to the semantic and pragmatic contents of statements, to collective assemblages of enunciation, to a whole micropolitics of the social field" (7).  The rhizome instead focuses on extending links to "diverse modes of coding (biological, political, economical, etc.) that bring into play not only different regimes of signs but also states of things of differing status" (7).  A rhizomatic study of language thus actively explores actual social spheres in which there is no single universal language, no homogeneity, but a complex interaction between all the articulatory modes, "not only linguistic, but also perceptive, mimetic, gestural, and cognitive" (7).

     Although Deleuze and Guattari note that there is "no mother tongue" but rather a "throng of dialects, patois, slangs, and specialized languages" (7), there is

a power takeover by a dominant language within a political multiplicity.  Language stabilizes around a parish, a bishopric, a capital.  It forms a bulb.  It evolves by subterranean stems and flows, along river valleys or train tracks; it spreads like a patch of oil.  (7)

The rhizomatic method attempts to take into account this multiplicity.  It differs from Chomsky's tree-based method in that it does not attempt to "break a language down into internal structural elements, an undertaking not fundamentally different from a search for roots", but instead forms an analysis by opening it to "other dimensions and other registers" (7-8).

     If this is the method of study that Deleuze and Guattari favour for analyzing language — a method that takes into account the nature of the subject matter — then it may also be possible to construct a rhizomatic analysis of the Net, constituted as it is by connections between diverse, heterogeneous spheres.  Barrett hazards a straightforward definition of the Net that can help us to see its rhizomatic form:

From a purely technical perspective, the Internet can be defined as the set of standards, protocols and mechanisms that allow a world-wide set of computers — from mainframe to PC — to interact, sharing software and data.  (1996: 12)

So even as a physical entity, the Net comprises connections linking points or nodes that are conspicuously diverse, 'speaking' diverse 'languages' — in technical terms these would be coding schemes and operating platforms.  Potentially, the diversity of the Net, then, is even greater than Deleuze and Guattari claim for natural languages.  These connections are essentially constituted by physical machines working in conjunction with TCP/IP, a common 'metalanguage' or protocol, channelling information through various media.4

     Opening up his definition of the Net, so that it is not confined to just the connecting technology, Barrett offers this more general meaning:

Internet: a global pool of information and services, accessible by means of locally executed interface software.  (1996: 12)

By doing this, he casts a wider net, ropes in a bigger, more nebulous and loosely defined creature.  Upon opening his net, he finds nodes of all descriptions: personal home pages, government bureaucracies, movie databases, library catalogues, e-mail discussions, commercial interests, non-profit organizations — all in different languages, something very like Deleuze and Guattari's "throng of dialects, patois, slangs, and specialized languages" (7).  Although Barrett does not refer to Deleuze and Guattari, his idea of the Net is a textbook example of the rhizome, with no apparent organizing structure, no determining point of origin, no tyranny of centrality.  Nevertheless, stabilizing points do dominate.  For example, the Net began in America (Hafner & Lyon, 1996), and the US dollar is the currency of choice for online trade, just as global exchange rates prefer to be wedded to the greenback.  Also, English dominates, although other languages reserve a fighting chance: "Language stabilizes around a parish, a bishopric, a capital" (7).  Barrett's version of the Net is indeed "a tuber agglomerating very diverse acts", and it "spreads like a patch of oil" (7).

     Historically, the principles of connection and heterogeneity were fulfilled by an important development of the Net.  In 1965, the Advanced Research Projects Agency (ARPA) received and approved a research proposal from a small and newly formed computing firm which was looking for contracts after losing a major investor (Hafner & Lyon, 1996: 68).  Tom Marill, who made the proposal to tie up two different computers over long distance, had a specific objective:

[C]omputing had reached an unfortunate state of affairs; time-sharing projects were proliferating, but there was no "common ground for exchange of programs, personnel, experience, or ideas."  His impression of the computer science community was "of a number of essentially similar projects, each going off in its own direction with complete disregard for the others."  Why waste resources?  (1996: 68)

As a result of the personal interest and obsession of a team of unorthodox scientists and engineers managed by Larry Roberts at ARPA, Marill's modest experiment led to the ARPANET project, which eventually blossomed into the Net.  Marill was not seeking to bring disparate technologies into anything like a tree-like convergence, but to tie them together to avoid duplication.  In effect, he was seeking to connect heterogeneous elements.


A multiplicity, according to Deleuze and Guattari, is an assemblage of points to be taken as a rhizomatic whole.  The one has no useful value except as part of the many.  In the multiplicity, "the multiple is effectively treated as a substantive" whose operations go beyond the politics of subject and object positions (8).  Instead, the multiplicity comprises "determinations, magnitudes, and dimensions that cannot increase in number without the multiplicity changing in nature" (8).  In other words, the multiplicity transforms itself, shifts in magnitude and quality ("the laws of combination therefore increase in number as the multiplicity grows"), never remains a static entity, as it extends links to and recruits ever more dimensions (8).  Deleuze and Guattari use the example of the multiplicity of puppet strings connected to the "multiplicity of nerve fibres" in the arms, which are in turn linked to the brain, the "undifferentiated" (8).  But the idea of the brain as "undifferentiated" can be pushed back still further by thinking of it as a multiplicity of synaptically interconnected neurons (Marshall & Zohar, 1997: 241).  Hence the multiplicity proliferates by extending connections to other multiplicities.

     Deleuze and Guattari say that a multiplicity is not composed of "points or positions" but "lines" (8).  Because a multiplicity does not have a "unity" like a "structure, tree or root" in which elements of the set are placed, it cannot be measured by a simple act of counting — some "notion of unity (unité) appears only when there is a power takeover in the multiplicity by the signifier or a corresponding subjectification proceeding", which is precisely an operation of a tree-dichotomy (8).5  That kind of binary logic imagines a "supplementary dimension" transcending that in which the system exists, "overcoding" or determining the system (8-9).  Therein lies the difference of multiplicities, for they are "flat", in the sense that they are only made up of the dimensions in which they reside, and nothing more (9).  Therefore multiplicities exist on a "plane of consistency" (9), or a plane which operates "a holding together of disparate elements" (Massumi, 1987a: xiv), a "grid" of connections which exists "outside of all multiplicities" (9).  Multiplicities are held together by this "outside" — the dimension-between-dimensions in which exists "the abstract line, the line of flight or deterritorialization according to which they [the multiplicities] change in nature and connect with all multiplicities" (9).  In other words, multiplicities are not structured by a transcendent, overcoding dimension, but are facilitated by an outside dimension-between-dimensions of connections.  A multiplicity can never be self-contained or "constituted by the interiority of a substance or subject" (9); rather, it is always part of some wider picture.  As such, the boundaries of a multiplicity can never be clearly drawn.

     How then is the Net a multiplicity?  Firstly, it cannot be measured: "Nobody knows how many people use the Internet, because ... it is a network of networks" (Negroponte, 1996: 181).  Secondly, it does not sit well as a set of countable elements to begin with — for it is not a one; it is not even a many, but a one composed of many.  Thirdly, many computers need to be connected before they become a multiplicity.  The Net is just such a collective entity, a whole made up of a multiplicity of elements: software, hardware, humans, personal computers, mainframes, routers, cables, modems, phone lines, Telnet, World Wide Web, FTP, credit card databases, digital cameras, ATMs, stock exchange listings, publishing houses, virtual companies ... all multiplicities within multiplicities.

     If we ignore all the multiplicities of the Net except one — the grid of connecting equipment (phone lines, cables, routers, and so on) — what remains is a plane of consistency, an outside dimension-within-dimensions, holding the Net together.  This is not static, but changes in size and nature with every new connection, and every new type of connection, made.  New dimensions are always being recruited: yesterday, on-line libraries; today, electronic commerce; tomorrow, perhaps virtual reality.  The Net is a dynamic flux of multiplicities that resists overcoding and determination.  Yet there are still points of transcendental determination on the Net, which we shall discuss towards the end of the chapter.

Asignifying rupture

To understand this rhizomatic principle, we must first understand what Deleuze and Guattari refer to as "becoming" — a notion served by the processes of deterritorialization and reterritorialization which are "always connected, caught up in one another" (10).  If deterritorialization ruptures the line delimiting the conceptual or physical space of the subject, then reterritorialization redraws that line — not necessarily restoring it to what it was, but clearly redefining the subjective space.  We can be assured that the first line of demarcation is merely a transient state, a lamp-post passing by in the scenery from a moving train.  To use another metaphor, Deleuze and Guattari illustrate these processes using the example of the pollination of an orchid by a wasp.

     The orchid deterritorializes its 'plant-ness' or 'flower-ness' by rendering itself in the image or tracing of the wasp.  The wasp then makes a claim on the image, reterritorializing it.  Simultaneously, the wasp is deterritorialized by the orchid, "becoming a piece in the orchid's reproductive apparatus" (10).  The simultaneity of these two processes is not a linear operation of signification ("mimesis, mimicry, lure, etc.") in which one imitates the other in one direction only, but something beyond a "parallelism between two [different] strata" (10).  The strata of signification themselves have been ruptured:

[S]omething else entirely is going on: not imitation at all but a capture of code, surplus value of code, an increase in valence, a veritable becoming, a becoming-wasp of the orchid and a becoming-orchid of the wasp.  Each of these becomings brings about the deterritorialization of one term and the reterritorialization of the other; the two becomings interlink and form relays in a circulation of intensities pushing the deterritorialization ever further.  (10)

The two heterogeneous elements are linked then by a line of flight, a rupture between them which cannot be determined by signifying strata, but which form a rhizome.

     The concept of becoming is a paradoxical one: when A effects a becoming-B, A does not give up being A.  It continues to be A, yet it becomes B without transforming into B itself.  Deterritorialization cannot be separated from reterritorialization.  When Deleuze says, "Alice becomes larger", he means she is "larger now", "smaller before" (1993: 39).  But the moment now is so short that the moment before is indistinguishable from it.  The moment in which Alice was smaller and the moment in which she is larger are a part of the same simultaneity:

Insofar as it eludes the present, becoming does not tolerate the separation or the distinction of before and after, or of past and future.  It pertains to the essence of becoming to move and to pull in both directions at once: Alice does not grow without shrinking, and vice versa.  (1993: 39)

     What the concept of becoming, with its simultaneous deterritorialization and reterritorialization, amounts to is a blurring or reconfiguring of signifying distinctions — a common motif to be identified in information technology (IT).  One of the most obvious reconfigurations is to be found in the decreasing importance of what Ohmae calls the "historical accident" of geo-political demarcations in the context of IT-driven regional economies straddling the boundaries between nation states (1996: 5).  The distinction between the public sphere and private space is also becoming blurred — witness for instance Usenet discussions in the public sphere conducted from the private space of the home.  The phenomenon of telecommuting has also breached the line between work and leisure along spatio-temporal axes.  But most fundamentally, the Net is a jumbled hybrid of several earlier technologies, including the typewriter, the telegraph, Boolean logic, telephony, the cathode ray tube, transcontinental cable links, satellite-broadcasting, cryptology, the transistor, printed circuits, computing, and so on.  In this context, Rheingold writes: "Again and again, the most important parts of the Net piggybacked on technologies that were created for very different purposes" (1994: 67).  These technologies have all lost a degree of signification in their own terms, but through deterritorialization and reterritorialization (becoming the Net), they form lines of flight with each other.  At the same time, the Net could not have 'become' if there had not been asignifying ruptures between distinct fields of study — telecommunications, computing, psychology, military defence, and so on.6

Cartography and decalcomania

These two principles form departures from the vertical-branching form of the tree:

It is our view that genetic axis and profound structure are above all infinitely reproducible principles of tracing.  All of tree logic is a logic of tracing and reproduction.  ...  It consists of tracing, on the basis of an overcoding structure or supporting axis, something that comes ready-made.  The tree articulates and hierarchizes tracings; tracings are like the leaves of a tree.  (12)

It is this vertical dependence on the support of what came previously, this dependence on the genetic axis or fundamental blueprint for any form of branching outgrowth, to which Boundas refers as the "foul odors of inbreeding" (1993: 1).  The tree method capitalizes on and takes for granted such intrinsic tracings to find a way either up or down the structure, but all movements will always be trapped within the structure and be part of it.

     By contrast, rhizomatic links are not tracings.  They are formed through mapping — or active construction based on flexible and functional experimentation, requiring and capitalizing on feedback.  The map is not an image from which reality is to be traced (a reversal of photography which traces the image from reality), or a blueprint whose workability has to be taken on faith; the map is never fixed, but a changing flux of adaptation and negotiation.  It is intimately and mutually tied to all the other principles of the rhizome.

The map is open and connectable in all of its dimensions; it is attachable, reversible, susceptible to constant modification.  It can be torn, reversed, adapted to any kind of mounting, reworked by an individual, group, or social formation.  ...  The map has to do with performance, whereas the tracing always involves an alleged "competence".  (12-13)

And whereas the tree only has its genetic axis as the singular point of entry "upon which all successive stages are organized", the rhizome has "multiple entryways" (12).

     The history of the Net reflects this principle of mapping.  It is a common myth that the Net had one single point of origin which can be traced back to an intentional military project to survive a nuclear disaster (Hafner & Lyon, 1996: 10).  Rheingold describes the Net as having something of an "accidental history" (1994: 65), or rather a history of many happy accidents.  Although the credit for developing the Net is generally attributed to ARPA and the National Science Foundation (NSF) (Rheingold, 1994; Hafner & Lyon, 1996; Barrett, 1996: 21-22), a close investigation back to its earliest days in the 1950s reveals that there was never a deliberate plan for its construction.  Nobody came up with an architectural blueprint and announced, "Okay guys, let's build this thing.  Just follow the design and we'll have the Internet".7  Instead, the Net emerged gradually over the decades as a series of active negotiations and improvisations involving various interests, and eventually became what it is today.  Furthermore, the Net has not 'arrived' in the sense that it has reached a preordained state of completion.  Barrett has probably avoided a definition for the Net based on technical delineation (1996: 12) because the Net has not stopped evolving yet, and perhaps never will.  A better definition for the Net might therefore include mention of this constant shift in size and nature, which is precisely the rhizomatic attribute of mapping, as opposed to tracing.

     Insofar as the Net does not rely on any overcoding blueprint but instead depends on flexible adaptation and negotiation for its continued formation, it may be considered recalcitrant, not unlike grass or weeds.  Deleuze and Guattari quote Henry Miller:

China is the weed in the human cabbage patch.  ...  Eventually the weed gets the upper hand.  Eventually things fall back into a state of China.  ...  Grass is the only way out.  ...  The weed exists only to fill the waste spaces left by cultivated areas.  It grows between, among other things.  (Quoted in 18-19)

According to Hafner and Lyon, the Internet evolved from an experiment in 1967 called the ARPANET (1996: 68-81) — a computer network which essentially came up like grass or weeds when nobody was looking.  In the context of Cold War tensions during which the Eisenhower administration placed a high degree of faith in the scientific community, and at the height of the space race in 1958, ARPA was founded with a mission to develop defence projects in the style of Reagan's later 'Star Wars' programme, but was left high and dry when NASA made its appearance and usurped that role (1996: 20-22).  With its existence in crisis, ARPA redefined its purpose, dropping military defence in favour of science, and became a sponsor of long-term advanced research and development (1996: 22).  It was in this environment where researchers were given free rein — especially in the time of ARPA's third director Jack Ruina, with his relaxed management style and trust in researchers to "pick the best technology" (1996: 23) — that the Internet-weed took root.

     Indeed, the Internet-weed was not made welcome in the more cultivated military and commercial hotbeds.  In 1965, Paul Baran in the US responded to the popularly perceived threat of nuclear war with the idea of using redundant links to make a communications network less vulnerable to complete breakdown (1996: 56).8  Shortly after this, Donald Davies in Britain independently conceived a similar network based on redundant links; but whereas Davies was successful in obtaining funding for an experiment in computer networking, Baran received too little commercial and military support to realize his ideas (1996: 52-67).9  For years, Baran's recommendations were either variously turned down or offered insufficiently competent commitment (1996: 52-67).  However, on meeting Larry Roberts in 1967, he assisted the ARPANET team on a casual basis, even though Roberts was not too concerned with building a network that would survive a nuclear disaster (1996: 77).  The point is that the Net was never a deliberately planned government or business programme; it could only have grown up in between these cultivated patches.

     Functionally, the Net also works on active mapping.  Marill's networking experiment in 1965 involved a method of exchanging data between computers which he called a "protocol" (1996: 69).  This first protocol would be succeeded by improvisations still in use today, namely Telnet, FTP, TCP/IP and SMTP (1996: 174, 224, 236-237, 252).10  These protocols, which enabled and refined the connections between nodes, were adopted after practical and functional experiments on the Net; they were active constructions based on a process of continual mapping, not blueprints or tracings.  In 1988, the International Organization for Standardization — which had played no prior role in building the Net — declared another protocol, the Open-Systems Interconnection (OSI), the official standard to replace TCP/IP.  Although OSI was embraced in bureaucratic circles, it never caught on.

OSI had a more complicated and compartmentalized design.  And it was a design, never tried.  As far as the Internet crowd was concerned, they had actually implemented TCP/IP several times over, whereas the OSI model had never been put on the tests of daily use, and trial and error.  (1996: 247)

To borrow the words of Deleuze and Guattari, TCP/IP is the map that has to do with "performance", whereas OSI is the tracing that involves an alleged "competence".  TCP/IP as a product of trial and error has to be a connectable mapping, adaptable and open to functional modifications centred on performance, "an experimentation in contact with the real" (12).  OSI on the other hand was conceived in isolation from actual implementation, always returning to the original genetic axis of the engineered blueprint.

     Also noteworthy in this respect is the fact that the OSI design was rooted in the centralized effort of one organization.  On the other hand, TCP/IP and its related protocols were the result of the ad hoc efforts of various programmers who actively sought feedback from Net users through a channel called Requests For Comments (RFCs) which is the "accepted way of recommending, reviewing and adopting new technical standards" and "a simple mechanism for distributing documentation open to anyone", used for disseminating ideas and "spreading the network culture" (Hafner & Lyon, 1996: 145).  In this way, RFCs serve as a channel for feedback in the active mapping and territorialization of the TCP/IP rhizome.

     TCP/IP itself functions as a rhizomatic working between two distinct protocols, TCP and IP, but it also has the ability to easily form rhizomatic connections.11  Its success as the de facto protocol was largely due to its free distribution in 1982 as part of Sun Microsystems' new version of the UNIX operating system (1996: 250).  In effect, TCP/IP proliferated through a territorializing becoming with UNIX.

Just how rhizomatic is the Net?

So far, the six principles of the rhizome have been successively described in the Net.  Such a mapping of the rhizome to Internet technology, however, is not entirely new. Burnett has made the connection between the rhizome and hypertext (1993), and Hamman has applied rhizomatic principles to the Net as a whole (1996).  Both of these explore the positive analogy between the rhizome and Net technology.  What remains to be done is to find negative analogies in order to complete the picture.

     Here, we need to distinguish between the terms "rhizomatic" and "rhizomorphic".  "Rhizomatic" describes a system which adheres to all six principles of the rhizome, with no characteristics of the tree.  "Rhizomorphic" means having rhizomatic tendencies without excluding arborescent attributes.  I want to argue that the Net is not an uncomplicated instance of a rhizomatic system but rather a rhizomorphic one.  In other words, I shall argue that it is problematic to say that the Net is a rhizome, not a tree.

     To return to a previous example, TCP/IP as the haemoglobin of the Net has been described as being "unerringly open" (Hafner & Lyon, 1996: 250), able to link the Net with any computer or network over any form of connection.  It was this openness of TCP/IP which allowed various networks to connect with ARPANET in the 1980s, resulting in a vast network of networks eventually called the Internet (1996: 244).  Every time an existing network is connected to the Internet via a gateway, the multiplicity grows.  Hafner and Lyon refer to a "gateway" as a "routing computer" acting as a link between two networks (1996: 223), but the current usage of the term connotes a point of entry to an Internet backbone.  A backbone is a high-capacity connection analogous to a highway linking major cities around the globe.  Anyone who wants to be an Internet service provider (ISP) must first obtain gateway access to a backbone.  Once established, an ISP acts as a 'gateway' (entry point) for its subscribers to access the Net.  As we shall see shortly, this structure can suggest the rhizomorphic rather than the rhizomatic.

     Rafaeli's statement that the Net "treats censorship as noise and is designed to work around it" (Newhagen & Rafaeli, 1996: 5) recalls Baran's idea of redundancy as a way to increase the reliability of a network against complete communications breakdown.  With redundant links, when one particular route to a destination is choked, data can always find another way.  However, any survey of worldwide backbone routes will reveal that the level of redundancy is at present rather limited and not evenly distributed throughout the world.  Anecdotal evidence of notoriously slow connections to certain parts of the globe is so common that a vernacular term has sprung up: the "World Wide Wait".  Based on current trends, Stephenson has observed that by the late 1990s, the United States will be about ten times better connected to Europe and Asia than Europe and Asia to each other (1996: 109).12  Because redundancy costs money to build, economic concerns will always put the brakes on how far the principle of connection can be taken; less redundancy means less robust connections.

     Even if some level of redundancy exists between the backbone nodes on the Net, a breakdown can still occur at any point between the gateway and the individual user, where zero redundancy and several points of vulnerability exist (Garfinkel, 1996).  Most subscribers have experienced some instance of access interruption due to such technical incidents.  A rhizome, it is true, has multiple entry points; but if a particular gateway is the only one available in a geographic region, then there is effectively only one entry point.  I can always dial up another ISP, but if the nearest alternative involves expensive telephone charges, then I might as well wait for my access to come back on, or even do without.  Entry points should not be defined by their physical existence alone, but also by the ease with which they might be attained.  TCP/IP, then, may enable multiple entry points and connections, but it will not guarantee multiple entry points for access.  Indeed, in some less developed regions, Internet presence may be out of the question altogether — and certain social strata may be excluded from the Net for economic and other reasons.  This exclusivity puts a constraint on the level of heterogeneity of the Net.  To this extent, it cannot be considered fully rhizomatic.

     Moreover, a recent widespread outage has also drawn attention to the Net's potential for arborification, or returning to a tree-like form.  A corrupted "master list of top-level domain names (the .com that directs e-mail messages and Web browsers to the correct sites)" was mistakenly sent out across the Net by Network Solutions, which manages domain names in the US, and this "threw nearly the entire Internet into absolute disarray" for four hours before the problem was fixed (Rushkoff, 1997).  Network Solutions runs the profitable InterNIC domain registry and has some vested interest in centralizing and controlling Net traffic.13  Such commercial interests prefer an Internet-tree, which can be chopped down at the trunk, over an Internet-rhizome.

     As Deleuze and Guattari write: "Plants with roots or radicles may be rhizomorphic ... : the question is whether plant life in its specificity is not entirely rhizomatic" (6); and "[t]here exist tree or root structures in rhizomes; conversely, a tree branch or root division may begin to burgeon into a rhizome" (15).14  And so a tree may be a tree, but a forest is a rhizome; a computer may have a centrally structured memory and processor (16), but a network of computers can be decentralized and exchange information based on a map.  This map or "graph regulating the circulation of information is in a way the opposite of the hierarchical graph" (Rosenstiehl & Petitot, 1974, quoted in 17), but it is still possible to arborify the workings of the graph, or in the case of the Net, the domain list, to bring it to one central point.15  Accordingly, Deleuze and Guattari concede:

You may make a rupture, draw a line of flight, yet there is still a danger that you will reencounter organizations that restratify everything, formations that restore power to a signifier, attributions that reconstitute a subject.  (9)


The important point is that root-tree and canal-rhizome are not two opposed models: the first operates as a transcendent model and tracing, even if it engenders its own escapes; the second operates as an immanent process that overturns the model and outlines a map, even if it constitutes its own hierarchies, even if it gives rise to a despotic channel.  (20)

So the question remains whether the Net is a root-tree with rhizomatic tendencies or a rhizome with arborescent inclinations.  This may only be a matter of semantics; a more useful distinction between tree and rhizome may be that each is in its own way "a mode of calculation applied to things" (17).  What is important is that the Net has the flexibility to negotiate a territorialization with and between the two — "the same thing is generally susceptible to both modes of calculation or both types of regulation, but not without undergoing a change in state" (17).  What is this change in state?  The simple answer would be a shift between tree and rhizome, but let us examine other possible paradigms in the following chapters.

Chapter 2