How the language baked into the foundation of computing shapes the internet…
Deleuze and Guattari defined the rhizome as a challenge to the root-tree structure epidemic in critical thought. Since this seminal definition, the ontological structure (or lack thereof) of the rhizome has been readily applied to the internet. In its comparisons to a rhizome, the internet has been identified as an assemblage of connections that defy the problematic binarism inherent in the root-tree structure. However, challenges to this application of the rhizome are prominent. In this essay, I will explore just one of these challenges: examining the hegemonic nature of the English language as it applies to the computer, the basic building block of the internet. First, I will outline the privileged position of the English language in the basic architecture of the components of the internet, from the machine level to the level of programming language and internet protocol. Then, I will argue that this contradicts Deleuze and Guattari’s basic edict that a rhizome has no ‘hierarchical modes of communication and pre-established paths’ (1459), and that instead, this hegemonic power of English has formed the internet into only a partial rhizome, if not a solid structure constructed and continuously policed by discourse.
Koh observes that ‘the Net is a textbook example of the rhizome, with no apparent organizing structure, no determining point of origin, no tyranny of centrality’ (Koh); his claim is not a rare one. The internet’s structure as decentralised, as a system of interconnected nodes, is abundantly explained too by Nicholas Negroponte, who describes it as a ‘network of networks’ (181) whose form makes the shape and size of the whole unobservable. He goes on to espouse the idea that the internet is ‘a lattice of heterogeneous processors’, in the manner of somebody describing a rhizome without having read Deleuze and Guattari, and he even goes on to say that two parts of the lattice are ‘so totally different from each other that their designers don’t even speak the same language’ (180). Suhail claims that ‘within the digital network, communication is “organised” as a rhizome’, and expands on this to say that the internet is also rhizomatic in that it has ‘no main artery’. Rothstein supposes that the rhizomatic form of the internet is so inherent that the rhizome itself is the ‘symbol of the internet age’. With this language, it seems a given that the internet is a near-perfect rhizomatic form. In this essay, I aim to prove that this is not so, and to prove that this fact has wide-reaching implications for those exposed to nodes on the network.
The internet’s network is composed of centres; this is troubling for those who compare it to a rhizome. Of the rhizome, Deleuze and Guattari say: ‘it is not a multiple derived from the One, or to which One is added (n+1)’ (1458), but the internet is fundamentally composed of the sum of Ones. Rather than the rhizome that is ‘made only of lines’, the internet is very much a structure in that it is very much ‘defined by a set of points and positions’ (1458). Each node is a computer with a definitive boundary between the internal (the local storage paradigm) and the external (the network communication paradigm). This distinction is key because it dispels the notion that the node cannot be separated from the system as a whole. In fact, there is a very definite threshold between the inside of the node and the interconnected space between nodes. What I am most concerned with for now though, is the way in which each part of the computer node speaks (the internal and external), i.e. how each encode and decode communicative language.
The network nodes that constitute the internet are intolerant of any language that does not use the basic Latin alphabet.
First, I will deal with the internal language. At the most fundamental level, the individual node stores and manipulates data in binary, in a series of on-or-off states. To store any information from a set, the computer must group these binary ‘bits’, which can each represent one of two states, into larger units that can store more than two states. The early architects of the ancestors of the computers, used as network nodes today, chose to group bits into groups of eight, each group being known as a byte. A byte has 128 possible states, being composed of eight bits, each with a total of two possible states. Character encoding is the next layer towards representing language. The predominant character encoding system for many years was ASCII, which is able to represent a full set of characters with one byte per character, because there are fewer than 128 characters in the set: those of the Latin alphabet, the Arabic numeral system, and assorted punctuation and other symbols (‘ASA standard X3.4–1963’). Whilst this character set is able to compactly store and represent language, it is drastically limited and excludes the human user from interacting with any language other than that derived or representable in the Latin alphabet. Whilst new character sets have been invented that can store characters other than the basic set of Latin characters, there are none that are widely compatible at the low-level of operation, and widely inclusive of other languages or script systems. The details here are technical, but the consequences are that the network nodes that constitute the internet are intolerant of any language that does not use the basic Latin alphabet. We have seen then, that unlike the rhizome, the internet is a network certainly composed of a multitude of Ones. We have also seen that those nodes are deeply preferential in and of themselves, of a certain kind of language.
One thing has been well observed about the internet: that its contributors and architects do not reside in a single geographical location, nor do they exist in a centrally administrated structure. By having agreed upon a set of shared protocols for communication and information markup (TCP/IP, HTTP etc.), disparate and unorganised agents are able to develop platforms for information to flow across the network, as well as the information that flows across these platforms. All of this is done concurrently and without central coordination. However, the question remains: what kinds of information are tolerated by the network, and in what language can agents develop platforms for the network? The languages used to program systems for the internet (called simply, development languages from here on) use human language (that is, language as it is commonly understood: spoken and written language) fragments to represent the data manipulation and logic decisions being undertaken by the computer behind-the-scenes. No development language with non-trivial usage uses fragments from any human language other than English. Wherever they happen to be, geographically or on the network, when writing in development languages, architects must communicate with the computer in terms of ‘for loops’ and ‘if statements’, and always in those English-based terms specifically.
No matter where an internet node is, they are situated in the Anglophone world.
I also wish to refer to those languages that are more properly called protocols, the specific methods by which nodes exchange information packets (I am now speaking about the communicative language of the computer’s language for the external, the network communication paradigm). These are what Negroponte and Koh so optimistically call a ‘metalanguage’, allowing for an error-free exchange in diverse languages so long as the metalanguage’s protocol is adhered to (Koh). However, what is omitted here is that the metalanguage itself is an English-speaking one. A node requests a document from another node with the HTTP protocol using the keyword ‘GET’ and gives a document to another node using the keyword ‘PUT’ and so on. The ‘conspicuously diverse’ nodes that Koh refers to are in fact exchanging in development languages based on the English human language, all the while adhering strictly to an English-based metalanguage (Koh). In once sense at least, Negroponte was right when he claimed that the internet would herald a ‘space without place’ where the ‘post-information age will remove the limitations of geography’ (164). However, that ‘space without place’ represents the fact that no matter where an internet node is, they are situated in the Anglophone world.
Here are the first signs of Foucauldian discourse at work; to determine what power relations are at work, we must ask ‘according to what rules has a particular statement been made, and consequently according to what rules could other similar statements be made?’ (‘Archeology of Knowledge’ 31). If all expression by the architects of internet services and platforms are constrained to be conceived in English, this surely represents a discursive rule, itself constitutive of a discourse. If we introduce such a dominant discourse based on the constraining of the internet into English, do we disrupt the rhizome? To speak in a spatial metaphor, I posit that discourse comes to centralise certain parts of the rhizome, thickening some pathways more than others, until the whole begins to resemble a root-tree structure. Koh notes that the internet ‘is a jumbled hybrid of several earlier technologies’ with disparate purposes in line with Deleuze and Guattari’s definition of a rhizome as containing asignifying nodes (Koh), but of course this does not close the door for discourse as ‘discourse must not be referred to the distant presence of the origin, but treated as and when it occurs’ (Foucault, ‘Archeology of Knowledge’ 28). Therefore it is different for me to argue that the underlying architecture of the internet leads to the maintenance of a discourse today, because it continues to reinforce the ‘rules’ of language as a consequence, than for Koh to argue the inverse simply because the apparatuses going into the construction of the internet were not built to enforce discourse. It is not the intention of the components of the internet that are constitutive of discourse in this context, but their effect as enforcers of language ‘rules’.
I have demonstrated how the internet’s supposedly rhizomatic ‘flatness’ is disrupted by the force of an English-speaking discourse, where all communication across the network is marshalled and administered by English-speaking hegemonic forces. This disruption is an inevitable one for some, like Jo Freeman, who claims that ‘there is no such thing a structureless group’ and that a group will ‘inevitably structure itself in some fashion’ (232). In this context, the internet represents that group, and Freeman’s observations are analogous for many reasons. Firstly, she argues that it is the very heterogeneity that those like Negroponte claim is the lifeblood of the internet that causes emergent hierarchy; where heterogenous nodes form factions to represent themselves. Freeman also offers us an explanation for why the internet has been thus far diagnosed as an untroubled rhizome: ‘structurelessness become[s] a way of masking power’. This is how the internet has been misinterpreted as a rhizome without hierarchy. In much the same vein as the criticism Hardt and Negri make of the postmodern critique, I claim that those who claim that the internet is not troubled by structuring power because of its ‘difference, fluidity and hybridity’ have failed to recognise the true pathways through which discourse exerts its power (2622). Discourse shapes the behaviour of every single node on the network, and does not emanate from any single origin, nor does it need to exist in any explicitly spatial or physical apparatus of hierarchy. Because each of the nodes of the internet is constructed to ‘speak English’ at their most basic level and thus speak to one another ‘in English’, the whole structure exhibits a power discourse without any central authority or architect.
Those rules come to form a discourse that permeates and orders the whole system, disrupting drastically any notion of a flat rhizome.
With the English-speaking discourse of the internet highlighted, and a rationale for its creation of an anti-rhizomatic structure, it is now worth exploring some empirical examples which show this discourse manifested. Specifically, I would like to draw attention to examples that show nodes on the rough end of the discourse being administered by those benefiting from the discourse, or rather those at the core of the newly arborescent network. To assess the skew in the internet network, consider the fact that, as of mid-2014, there were 2.9 billion internet users (about 38% of the world’s population), 45% of whom were in Asian countries, 11% of whom were in North America, and 9% of whom were in the United Kingdom (‘Demographics’). Anglophone nations elsewhere in the world represent a trivial proportion when compared to the swathes of users in the Asian region. Given that 45% of internet users at the very least speak a language other than English as their first, and that the centric powers of the English language only make up 29% of the internet’s users, how is it that 55% of the internet’s content is in English (‘Demographics’)? This is a demonstrative example of how the English language’s preferential treatment as the language of each node in the network coerces all that passess through it to adopt the English paradigm. It is of little surprise, given that those organisations that continue to administer the internet and its services today reside in North America. Google, whose search algorithms define for many what information is accessible across the network, are ‘globally’ headquartered in California, a fifteen minute drive away from Apple’s ‘global’ headquarters. Amazon, who host a significant portion of the world’s distributed servers, is headquartered in Seattle. This is the very ‘main artery’ that Suhail claims the internet lacks, one of its key qualifiers as a rhizome. The coalescing of the network in terms of infrastructure and hegemonic power (together, Amazon and Google define much of the web’s evolving protocol and content) are indicative of the node’s preference for English. This has constituted a discourse that concentrates power around those at the centre of the discourse, ie. at the centres of spoken English.
Thus far, I have showed how the popular notion that the internet is a rhizome is a flawed one. The internet is not a system made without any determinable fixed points, like the rhizome. In fact, it has physically fixed points that define the network. I have showed how each of those points has been built in a way so as impose ‘rules’ about what kinds of language can be used across the network they contribute to. Those rules come to form a discourse that permeates and orders the whole system, disrupting drastically any notion of a flat rhizome. The heterogeneity and distribution of the network is threatened by a force of discourse that concentrates power around the traditional sources of English-speaking hegemony in the West, specifically Anglophone Europe and North America, and specifically in the purview of organisations like Amazon and Google. These concentrations of nodes within the network come to form a ‘main artery’ in the rhizome, annihilating its status of a rhizome at all. What are we to do with this internet, that seems to be a reflection of the material conditions of its creation, complete with hegemonic forces that privilege some in concentrated pockets, and disadvantage the rest? There is not room to address that here, but suffice it to say, the internet is certainly not a plain rhizome.
This article was originally written on March 2nd 2015 as part of studies on Critical Theory at The University of Exeter.
Deleuze, Gilles and Félix Guattari. ‘A Thousand Plateaus: Capitalism and Schizophrenia.’ The Norton Anthology of Theory and Criticism. Ed. Vincent B. Leitch. London: Norton, 2010. 1454–1462. Print. Foucault, Michel. The Archeology of Knowledge. London: Routledge, 2002. Print. Foucault, Michel. ‘Discipline and Punish: The Birth of the Prison.’ The Norton Anthology of Theory and Criticism. Ed. Vincent B. Leitch. London: Norton, 2010. 1490–1502. Print. Hardt, Michael and Antonio Negri. ‘Empire.’ The Norton Anthology of Theory and Criticism. Ed. Vincent B. Leitch. London: Norton, 2010. 2621–2635. Print. Koh, Chuen-Ferng. Internet: Towards a Holistic Ontology. Murdoch University, November. 1997. Web. 25 Feb. 2015. Livingstone, Sonia. ‘Critical Debates in Internet Studies: Reflections on an emerging field.’ LSE Research Online, 30 Jan. 2007. Web. 25 Feb. 2015. Malik, Suhail. ‘Harvesting The Tubers: Is The Internet A Rhizome?’Mute, 10 Jan. 1997. Web. 25 Feb. 2015. Negroponte, Nicholas. Being Digital. London: Hodder & Stoughton, 1996. Print. Norris, Pippa. Digital Divide. Cambridge: Cambridge University Press, 2006. Print. Rothstein, Adam. ‘Making Internet Local.’ Rhizome, 6 Nov. 2014. Web. 25 Feb. 2015. ‘ASA standard X3.4–1963.’ American Standard Code for Information Interchange, n. d. Web. 25 Feb. 2015. ‘Demographics.’ European Travel Commission Digital Portal, n. d. Web. 25 Feb. 2015.