Literature After Language’s Algorithmic Normalisation: Spam, Code, and The Digitality of Print in Blood Rites of the Bourgeoisie

Scott Wark

Abstract

The development of new media has a double effect on media that already exist. Old media register the impact of new media through the development of new artistic techniques, whilst new media clarify the technical logics and limits of the media they supersede. This essay uses Stewart Home’s contemporary novel, Blood Rites of the Bourgeoisie, to examine the impact of widespread algorithmic language processing on literary experimentation. Home’s novel incorporates appropriated spam email into its body. By doing so, it not only registers the impact of this ubiquitous form of contemporary communication, it also invites, and problematises, code and software based analyses. This essay approaches Blood Rites as the terminal point for a network of globally distributed processes of spam generation. Whilst it uses tools from code and software studies to engage with this text, it also argues that Blood Rites - and the phenomenon of spam that it mobilises - exceeds the distinctions that code and software studies use to isolate their objects of analysis. Highlighting the distinction between code and language as the chief stumbling block to apprehending spam, this essay proposes an alternate conception of the code/language relation. It suggests that digitality is the operative logic of both code and language, and it uses this logic to understand spam at both micro and macro levels - and the effects of this media technical process on contemporary literary expression.

Literature in the contemporary media-technical system

The much-vaunted supersession of print by electronic formats has yet to materialise. Though print may be far from dead, it’s no longer the only relay literary language passes through on its way to being read. Language is also processed by more or less automated code-based systems that deform it in more or less intentional ways. Friedrich Kittler argues that the development of new media of inscription in the early twentieth century challenged and revealed print’s status as the “universal medium” (Kittler 1999: 5-6). Literary modernism registered the impact of these modes of inscription, such as photography, cinema, visual advertising, through experiments with language’s possibilities (Murphet 2009). Today, language’s manifestation in electronic formats is not as radical as it once was (as in e.g. Landow 1992). The alternate modes of navigation afforded by hyptertext, for instance, have been routinised by knowledge work and pathologised as agents of distraction (Hayles 2010). This kind of language processing is now banal. What becomes of literature after the proliferation of automated language processing?

This essay will use Stewart Home’s recent print novel, Blood Rites of the Bourgeoisie (Home 2010), to examine the effects of this media-technical situation on literary experimentation. This novel registers the impact of perhaps the most banal automated language of all: email spam. Home’s text is very funny, but its most interesting aspect is its method of composition: chunks of appropriated email spam are interspersed throughout its narrative sections. By acting as a spam delivery system, Blood Rites operates as a point of contact between the printed book and the global, non-anthropocentric and code-enabled process, spam. Home’s use of spam invites us to understand this novel as a technical assemblage rather than a narrative-driven text because it mobilises a network of human and nonhuman agents that are “irreducible” to distinct “objects” (Latour 1988: 159). To understand Home’s text, this article will use code and software studies to explain how spam works.

Previously denoting cryptological ciphers or systems of law (Kittler 2008: 41-2), code is now used as “a synonym [for] computer programming language or as a text written in such a language” (Cramer 2008: 172; Lessig 2006 contradicts this by arguing that “code is law”). David Berry argues, persuasively, that code and software studies are “two sides of the same coin” (Berry 2011: 32). For Berry, the study of code implies a medium that is read closely, whilst the study of software thinks of code as a thing and reads it from a distance (Berry 2011: 32). Both code and software studies can be used to link these close and distant levels that converge on Blood Rites. In the field of literary studies, Alan Liu has perceptively commented that the vast corpuses of material analysed in “distant reading” struggle to bridge or articulate the ambiguous “contact zone” between distant and close literary critical models (Liu 2010). The code/software distinction suffers from the same problem. Blood Rites is not reducible to singular spam unit or to spam generation process. What is needed to secure Blood Rites’ vertiginous scaling from spam to code to print is a “midlevel concept” (Frow 2010: 248-9) that connects the distant and the close.

Yet, Berry also suggests that the medium of code transforms how other media relate to one another (Berry 2011: 10). New media clarify the strengths and weaknesses of the formats they supersede. In Wolfgang Ernst’s terms, this transition allows us to “reverse engineer” old media’s operative logics (Ernst 2013: 55). This essay will argue that Blood Rites exploits this transitional insight, and that in doing so it clarifies a key operational logic linking electronic and analogue language: their digitality. Digitality will be used as the “midlevel concept” that bridges the levels of the distant and the close in Blood Rites, and to engage with the slippery phenomenon of spam, by applying its logics to the code/software divide. Digitality refigures the code/language and new/old media distinctions that organise code and software studies. This essay will stage a confrontation between a literary text and global computational processes by proceeding, in descending order, from distant to close analyses of Home’s text. To engage with a book like Blood Rites, the distinctions that order code’s/software’s study must be refigured - by using digitality to rebuild the paths from the closest level back to the distant. Digitality is, perhaps, the technical basis of spam’s irreducibility to either human non-meaningfulness or nonhuman machine readability.

Kittler has argued that “new media do not make old media obsolete” but rather “assign them other places in the system” (Kittler 1996; also 1990). This essay risks a speculative gambit: that an inert media object, a book, can be used to clarify the relationships between code and language, new media and old, and media technologies and literary experimentation. Spam - and, through its appropriation of spam email, Home’s text - betrays an accelerating (if asymptotic) tendency towards meaninglessness, or towards the total negation of literary style. Perhaps the horizon of literature after language processing’s banalisation is just this: style’s machine generated negation. Perhaps, then, the digitality of code can be used as a lens to help us understand the scaling and interlocking media systems of today - and literature’s place in them.

Exhausting the same old magic trick, or, How to avoid recuperation

Home has been an active underground author, artist and self-declared one-man avant-garde for decades. Blood Rites is a melange of new and old media formats. It’s is short - only 127 pages - and divided into three sections: an unconventional narrative; a fictional erotic exchange between a character and the author of a real-life call-girl confessional; and a lengthy appendix containing fictional and probably pseudonymous posts and related comments from Home’s blog. Though fascinating as a literary object of study, this essay is not interested in making sense of Blood Rites’ literary subversions (see Turner 2002, on Home the author). In the tradition of the Situationist International - Home’s study of them was one of the first available to the Anglophone world (Home 1988) - Blood Rites is constructed to resist such critical recuperations. Home’s book is consequently hard to analyse. This essay will approach it, instead, as media.

Blood Rites merges Home’s abiding interest in schlock and pulp stylistics with a lowbrow aesthetic native to the internet age. Home composed this novel by signing his email account up to spam advertisers, building a collection of attractive examples, and interspersing these throughout the fiction section of his book. He uses the Situationist technique of correction, or détournement (Debord and Wolman 2006), to inject these spam texts with a pace and rhythm that reproduces the distracted cognitive experience of scanning internet pages (Hayles 2010). Here’s an indicative sample:

Link to Elizabeth McAlpine’s private pictures (link missing). You look so familiar! Have we met? Hey, what’s up? I believe we got drunk together at a party last week. Are you the guy that took Elizabeth home? Elizabeth told me she had a wonderful time with you! Anyway Elizabeth just gets sexier and sexier. Here’s the link to Elizabeth’s private picture, PLEASE do not give this out to anyone, this is for you ONLY! Check out her private pictures NOW! (Home 2010: 11)

Why turn to code and software understand what Blood Rites does? Because spam is a fiendishly difficult phenomenon to pin down, cutting as it does across computational and non-computational media. We can’t apprehend what Blood Rites does without using code and software studies to apprehend spam.

Literary studies has engaged with code in various ways. It has emphasised the medium specificity of code-based literature (Hayles 2004). It has examined the poetic and syntactical splicing of code operators and language in codeworks (Cayley 2006; Raley 2002). It has argued that the code can be read as literary text (Marino 2006). It has linked literary aesthetics to computational forms, like databases (Hayles 2012; Liu 2004). On a grand scale, it has used computational tools to refashion the way literature is studied and conceptualised (Moretti 2007; Liu 2010). As Florian Cramer points out, code-based studies of literature are usually “fixated” on literature that manifests “as self-contained works” modelled on belles lettres: formats like the webpage or file (Cramer 2013: 231-2). Rarely does the literary study of code engage with linguistic processes. This is precisely the kind of engagement that Blood Rites invites.

Partially, this is because Blood Rites uses spam to avoid making sense. As a phenomenon, spam pushes language to its outermost limit - to the point at which language no longer means (Galloway and Thacker 2009: 254). Home’s book is simply hard to read: these chunks of dètourned text direct spam’s assault on meaning, that implicit horizon of literary enquiry, at literature itself. Blood Rites is purpose-built to resist the literary critical urge to discover and disclose what it’s really about (a tendency Home is aware of: see McCarthy and Home 2001). Home uses his main protagonist, Suicide Kid, to announce the “point” of this book about mid-way through. We are told that Blood Rites is:

[drawing] attention to the way in which [women’s] work is devalued by a patriarchal capitalist society that views art activities and more specifically the role of the artist, as one of the few legitimate areas of male emotionality (Home 2010: 62).

It accomplishes this by inserting the names of famous female and queer male artists into détourned erotic email spam and by pummelling us with variations on this theme. By foreclosing the literary critic’s answer to the question, What does it mean?, this tactic frustrates the critical project’s usual prestidigitatory routine. There’s no rabbit for the critical reader to pull from their hat, because this rabbit is in Home’s hands. Nor, necessarily, are there iron clad “matters of fact” - or truths (Latour 2004: 237-8) - at stake in this book. These truths are right there for us to see, again and again and again. Another chunk of spam, another rabbit. Repeated to the point of exhaustion, the self-revealing sense of this book de -composes. What is brought to light is what is already there, so to speak. This book becomes a spam delivery system.

How can we come to grips with this book? Bernhard Siegert identifies a possible direction: by avoiding the “grand epistemic distinction between culture and technology, sense and nonsense, code and thing”, and by employing a “processual” approach to the assemblages of techniques and technologies that make up the media we interact with (Siegert 2013: 60). Real magic lies in the mechanics of the magic trick itself. Rather than critique Blood Rites by making it make sense, I suggest that we consider it as a “compositionist” nexus (Latour 2010) that mediates between textual and global-computational regimes of language in the act of spamming us, its audience. At the same time, Blood Rites reflects what literary experimentation might be in the age of globalised and banalised language processing.

To engage with the spam Home employs, code must be understood as one element in a larger, nonhuman media-technical network that brings the products of code-based processes “to the human” (Latour 1993: 82, emphasis added). But, a conceptual path, our “midlevel concept”, needs to articulate the local instantiation of this spam delivery system with spam’s global processes of generation if we are to understand what Blood Rites does. First, by exploring spam at the macro, or distant, level at which it is generated; second, by interrogating the relationship between spam and meaning up close; third, by establishing links between these levels. In the process, the distinctions that found code and software studies must also be questioned.

Massively distributed language processing: reading spam from a distance

The study of code begins with a definition of what code is. Like any object of analysis, code relies on formative definitions to mark it off from other phenomena. Two concern us here: that code is distinct from natural language, and that code operates in new media. When probed, these crucial distinctions concatenate into an open chain of interlinked divisions that beg the question: where exactly does code do its work? Is code a process of coding or is it an object that processes something else? Does code end with a line of code? A program? A computational system? The uses to which a computational system are put? The cultural practices that emerge from these uses? The embedding of these particular processes in the culture at large? - and so on. Each time it is decided that this is the degree to which code is open to its environment, boundaries are delimited; but, code studies’ degree zero is the idea that code is a distinct cultural phenomenon that requires specific modes of analysis. This calls code studies’ appropriate object of analysis in to question. Another way of putting this is to say that code slips between doing and meaning, between scales and mediums, or between code and software.

Software studies probes the “multiscalar reality” of code’s operations (Fuller 2008: 5) as these unfold in “social, technical, and aesthetic relation[s]” in culture writ large (Fuller 2003: 63; also, Manovich 2013). David Berry argues, persuasively, that studies of code and studies of software are “two sides of the same coin”: where “code is the static textual form of software”, software would be “the processual operating form”, such as specific software products (Berry 2011: 32; contra Manovich 2013: 16). Stated differently, the study of code is a kind of close reading, whilst the study of software reads the operations of code from a distance. Understood in this way, code is construed as a “supermedium” that radically reshapes how other media relate to one another (Berry 2011: 10; see also Manovich 2013: 105-6; Kittler 1999; 1996). What I would like to develop in this essay is how this supermedium is articulated between the close and the distant, between natural language and code, between new media and old. As I mentioned earlier, Liu has argued that the “contact zone” between the close and the distant is ambiguous. Might there be a “midlevel” concept that could avoid the blinkered specificity of the close and the fuzzy generality of the distant simultaneously (see Frow 2010: 248-9)? Can the epistemic distinction Siegert identifies between code and thing be avoided? In answering the question, What is spam?, we can scrutinise the distinctions set up by code studies. I will start with spam’s macro-level, its generation.

As a ubiquitous element of contemporary networked communication, spam’s familiarity makes it both easy to recognise and hard to categorise. On a very basic level spam is an irrelevant or unwanted message transmitted indiscriminately through a media channel. It is often sent to deceive its receiver into performing an action that will benefit its sender: defrauding users, infecting computers with viruses, or generating revenue through click-through advertising (Galloway and Thacker 2007: 146). Spam is heterogeneous: the kind that concerns us here is email spam, its most ubiquitous manifestation. We usually experience spam through singular examples: spam-units. Spam is not understandable as the simple sum of spam-units, though, because this reduction abstracts spam from the complex networks of algorithms and human and nonhuman actors that produce it. Spam introduces a problem of scaling that frustrates the levels of analysis implied by the code/software divide.

In his study of spam, Finn Brunton cites data showing that roughly eighty-five per cent of unfiltered email traffic is spam email (Brunton 2013: 153). This email traffic is filtered by systems put in place by administrators. These systems were originally created using naive Bayesian filtration techniques, which weight words based on their probable usage in legitimate or spam emails and which persist (in a heavily modified form) in the filters in use today (Brunton 2013: 140). Such filters work heuristically by learning what “normal usage” is: their rules are built by “reading” email contents and comparing emails marked spam with those marked legitimate. Filtration has spawned counter-algorithms calibrated to beat it in one of two ways: by replicating “normal” usage, or by spamming filters with enough “normal” content that regular messages are filtered and spam gets through (Brunton 2013: 150-152; Galloway and Thacker 2009: 254). Spam also exploits the “grey publicness” of network connectivity (Parikka 2013: 14). “Spiders”, or automated algorithms, can “scrape” addresses in bulk by exploiting networks’ reliance on call-response (“pinging”) feedback loops (Galloway and Thacker 2007: 146). From the point of view of the spammer, location - or rather, locatability - is more important than spam content. Increasingly, moreover, spam creation is automated in gigantic “botnets” that infect computers with viruses and link them together to exploit vast pools of idle resources (Brunton 2013: 199-200). Spam operates on a global scale and with such persistence because it is intimately related to the protocological affordances of our communication networks (Galloway and Thacker 2007: 145-7).

“Continuous paths”, as Bruno Latour put it, “lead from the local to the global” (Latour 1993: 117); if, that is, we see a phenomenon like spam as the product of a network of human and nonhuman processes. Spam scales very quickly from the spam unit to spam filters to massively distributed global networks inundated with unwanted data. Human intentionality intersects with nonhuman agency at each level. Machine learning systems shape our communication on a global scale as modulated by constantly shifting distinctions between acceptable and unacceptable communication (see Paasonen 2009). But, these distinctions aren’t semantic: they’re statistical, based on the patterns created by email usage. Filtered internet communication is statistically normative, its deviations statistically derived. Taken individually, spam filters produce an “everyday banality” (Fuller and Goffrey 2012: 14) of junk folders and improbable entreaties. Taken en masse, spam filters shape internet based language use. It doesn’t matter if you delete that spam email. The spam reached your inbox: that’s already a breach, a statistical loss.

Blood Rites levels a spam-laden broadside at distinctions between code and language, old and new media, mediums and relational processes of mediation, the human and nonhuman. Home jokes in Blood Rites that “literature is a Nigerian money scam”, because “cut and paste” “post-modern nonsense” uses the same compositional techniques as spam (Home 2010: 35). The force of this joke is lost if we apprehend Blood Rites on the spam-unit level. I’d like to suggest that Blood Rites should be understood as a spam delivery system, and that it can only be understood in relation to computational media and to the code on which these media run. Apprehended on its own, it’s just a particularly hard to read piece of literature that neuters the critical project in advance. To read Blood Rites is to be spammed.

This leads us to ask what spam is, exactly, and how it fits in to code/software distinctions. Is spam a code-based linguistic medium? No; it is generated by complex algorithms, but these algorithms are both multiple and inextricable from the communications they learn from and filter. Is spam software? Manovich claims that software is a “layer that permeates all areas of contemporary societies” (2013: 15, italics removed). Whilst I’m very sympathetic to the way software studies folds the “formalised” world of software back into the “messy” world of the real (Fuller 2008: 5-6), I think its use of machine-based programs as a clarifying object of study isn’t useful in this instance. Our objects of study dictate our tools; but, our tools also to some degree co-constitute our objects of study (Stengers 2001: 21-2). Focusing on the “cultural practices” of software’s “employment and appropriation” (Cramer 2008: 173) requires a software object to focus on, and around which a culture coalesces. Spam isn’t exactly software, and the culture it composes can’t be taken for granted as a thing that unproblematically exists (Latour 2010: 473-4; 1993: 103-4). Its medium isn’t culture, but rather a media-technical process that passes in and out of code based environments, or old media and new - and thus beyond the ambit of the “processes” described by Noah Wardrip-Fruin (Wardrip-Fruin 2012: 163-4). To account for spam, I’d like to clear the ground for conceiving of (both code and natural) language as digital phenomena, by revising the code/language distinction.

Code up close: digitality as techno-logic

…Only after commands, axioms, or, to put it briefly, sentences had been converted into numbers were they as infinitely manipulable as numbers. End of literature, which is made up of sentences. (Kittler 1999: 247)

Language is a fraught subject in code and software studies. A debate has emerged in the past decade about how code-based languages should be understood: whether code does things or whether it makes meaning. I’d like to refigure this debate by identifying how Blood Rites refigures and bridges the scales at which meaning usually operates. To do so, I will try to reconceptualise the code/language distinction at work in code studies. To distinguish code from natural language, code studies must first understand what natural language is and how it operates. An understanding of language that has found some purchase in code studies is performativity, and its precursor, iterability, which Jacques Derrida developed to describe the plasticity of linguistic signs - and which code studies has used to explain text’s ability to be encoded. I will discuss the various ways iterability has been taken up in the study of code before arguing that its material and technical aspects need to be emphasised to understand code’s operations - and the way a book like Blood Rites scales from the close to the distant. In the process, I will identify digitality as the logic underpinning the technical operations of both code and natural language, dissolving the distinction between them along the way. Digitality can also be used, as I will argue, to identify the technical foundations of the banalisation of language processing registered by Home’s text.

Derrida’s critique of illocutionary language acts - language use that makes things happen - is at the core of recent debates about the code/language distinction. In brief, Derrida argues that the performance of illocutionary acts is never pure (Derrida 1988: 18) but is always a kind of “citation” or “iteration” that exploits the language-unit’s capacity to break with its context and to be “grafted” into others (9-10). Two distinct uses of this concept have emerged. The first uses Judith Butler’s reworking of iterability into “performativity” to argue that code is “the only language that is executable” (Galloway 2012: 70-71; see also Hayles 2005; Butler 1997). This approach contrasts code with natural language by arguing that code operates in self-enclosed machinic environments that avoid “natural” illocution’s “complex chains of mediation” - the circuitous “external” routes of culture (Hayles 2005: 49-50). For Adrian Mackenzie and Theo Vurdubakis, this account is problematic because it erroneously separates “saying from doing” (Mackenzie and Vurdubakis 2011: 6-7). They argue that the performativity of code is related to the coding of the performance of actions by culture. Harking back to Derrida (1988: 17-18), meanwhile, Wendy Hui Kyong Chun argues that conceiving of code as executable succumbs to the fantasy of a self-present subject, the programmer, who enacts their intentions through code (Chun 2011: 28). For Mackenzie and Vurdubakis and for Chun, culture is the medium for code. What and where could code be? The question of what code is is inseparable, as we can see, from the question of where it works.

Another fissure separates these two conceptions: whether or not code makes meaning. Derrida argues that the meaning of language is not reducible to the intentions of its (algorithmic or human) author (1988: 9-11). But iterability also undergirds the promiscuous meaningfulness of language systems. For Galloway, code is symbolic but its capacity to signify is a secondary or “epiphenomenal” feature (Galloway 2010: 290; see also Galloway and Thacker 2009: 255). For Mackenzie and Vurdubakis, neither code and language nor “meaning and force” can be separated cleanly (2011: 7). Chun states this more forcefully by arguing that “[r]egardless of whether or not it can execute, code can be - must be - worked into something meaningful” (2011: 52). The critical code studies approach developed by Mark Marino radicalises this understanding by interpreting code’s text as meaningful, whether it’s meant to mean or not (Marino 2006). Spam is an interesting test case for these conceptions. Spam is polymorphous: the word can be used to denote a thing, an unwanted message; or an action, this message’s transmission. What becomes of meaningful language when language is processed by globalised spam filtering systems?

Meaning can be mined from all sorts of elements in language, including from the way words are put together - that is, from style. Brunton discusses the transformation of Neil Stephenson’s Cryptonomicon into indiscriminately targeted spam content by pranksters, quoting Stephenson’s fear that this procedure will “inoculat[e]” the world’s “spam-filtering systems” against his “literary style” (Brunton 2013: 147). Style is a signature that spam filtration systems can recognise, because style relies on repetitive tics that can be numerically weighted by sophisticated spam filters. Style becomes pattern. Filtration systems assemble stylistic norms heuristically. To them, distinctive style is measurable in statistical terms as degrees of abnormality. Style pushed to its limits in literary experimentation is a liability in machine-filtered communication.

Style is also practiced. Home embraces the style of spam - the attempt to reach the literary null-point of zero style, or algorithmically-verified linguistic neutrality - by détourning spam in his book. But, spam is not supposed to mean. Its meaning, if it has any, is usually a ruse or container that entreats its reader to take an action. Spam is never discrete; rather, it is a vector that exploits network relationality. What does all of this mean? That reading a spam email or Blood Rites for its message misconstrues spam’s significance. Like spam, Blood Rites is the ossified residue of an algorithmic battle for attention; only, it is reactivated once more by the reader-book relation. The spam Home détournes doesn’t stop working. It attests to a kind of techno-logic underpinning the translatability of email spam between massively distributed code-based environments and discrete mediatic instantiations. As Ernst argues, code-based writing is not simply “symbolic”: as a kind of writing it is also “eloctrophysical”, which is to say material (Ernst 2012: 61; see also Introna 2011: 116). This book invites us, I think, to get to grips with “the non-hermeneutic non-sense” that is the “base and abyss of meaning” (Siegert 2013: 52).

When we discuss code as a “neighbourhood of relations”, as Mackenzie and Chun advocate (Mackenzie 2006: 169; Chun 2011: 6), I suggest that we think these relations in material and technical terms. In saying this, I am suggesting that we complicate the idea that culture is the medium for code by co-implicating code and technology. Code’s iterability is afforded by technologies and techniques of inscription. Every writing or saying is a non-trivial rewriting or resaying. Code, which relies on data storage for its operations, is also inscriptive: “magnetic recording” is a kind of writing, for example (Kirschenbaum 2008: 29). Translation between media, including code, is non-trivial (Chun 2011: 23). In an alternate theoretical idiom, it has “costs” that must be accounted for (Latour 1988: 165, 1.2.5.2). If the “spacing” (Derrida 1988: 9; 1998, 68-9) on which iterability relies is understood in technical terms (Stiegler 2001: 249), this non-triviality can be expanded. Spacing refers to the physical distinctions between discrete graphemes that is integral to writing’s ability to make meaning. Technical media of inscription require supporting scaffolding - the page, for instance - for iterability to work. Code is more complicated again. Code does not break freely with its context, as iterability might suggest. Rather, it carries its context with it by defining the parameters that enable its execution in each and every instantiation (Galloway 2012: 64-5). Similarly, spam ties together heterogeneous chains of code-based and non-code-based media, including users. Spam can’t be understood if code-based language’s operations are limited to processing machines, or if culture is construed as code’s medium on the micro-scale of code or the macro-scale of software. This is not to say that “there is no software” (Kittler 1995). Rather, this is to acknowledge that these powerful conceptions of code are inadequate to the “quasi-objects” (Latour 1993: 108) assembled by spam. But, how does spam scale from the spam-unit back to the global process? I’d like to argue that digitality is the midlevel concept that affords this scaling.

As a materialised technique, digitality relies on the place-value system. The place-value system, or positional notation, is the function supplied by the numeral “0” in mathematics. This numeral has a spatial function. In the notation of “10”, the zero shifts the numeral “1” to give us “10”. This system is digital because it relies on the discrete numeral; but, this digitality is also inherently positional because it relies on the relations between digits. Like numbers inscribed language is also digital because it can be broken down into “discrete bit streams” (Hayles 2004: 76). Bernard Stiegler refers this as “discretisation”, the process of breaking continuous (analogue) flows into constituent, spatialised elements (2010: 10). Crucially, this process also relies on positional spatiality to operate. Meaning emerges from spatial arrays of letters. Because of this positional aspect, digital systems must “take into account the media employed to store and transmit them” (Siegert 2007: 38). They rely on spacing’s material manifestation. A book no less than a computer translates digital logic into the actuality of a technical object. This translation doesn’t stop with the manufactured computational or printed assemblage: it is embedded in their media-technical processes. Kittler:

Contrary to current opinion, codes are not a peculiarity of computer technology or genetic engineering; as sequences of signals over time they are part of every communications technology, every transmission medium. (2008: 40).

Digitality is a “genetic logic”, in Gilbert Simondon’s terms, that cuts across distinct media, that is supported by a material environment, and that enables computational and printed mediation (Simondon 1980: 60-65; Stiegler 1998: 68). The differences between code and language are not ontological but empirical, joining the book and computer in a recursive technicality irreducible to human history (Kittler 1999: esp. 231-263).

Spam is an example of a phenomenon that dissolves the technical medium, like an algorithm, into chains of mediation (see Galloway 2012: 17-18). Code and language are defined by the distinct media in which they operate. But, the digitality of both code and language is what allows these media to be interfaced and to form concatenating chains - networks - of purportedly distinct media forms. Digitality is the logic that transforms natural language into machine readable formats and back again. What matter here are not the ontological distinctions between code or language, but the work of media-technical manifestation - which Gilbert Simondon calls “concretisation” (Simondon 1980: 76-77) - that is the prerequisite of the work of machine reading or meaning making. Both kinds of work rely on the technical materiality of digitality. Interpretation or meaning production is a technical process, meaning a second-order effect (Kittler 1981). Natural language formats are only distinct for us, humans. From the point of view of machines, these formats are indistinguishable: what matters is their statistical content. Spam is not reducible to its either its form for us or for machines. Or, it is “irreducible” in Latour’s terms (Latour 1988: 157), and thus inextricable from the networks that afford it.

Bereft of any ability to trick us into a malicious mistake, the spam in Blood Rites breaks language down to its bare constituents: digits arrayed in space, without individual meaning, only working en masse but making less and less sense as they’re repeated. Its message can be over-coded with content (a critique of gender relations in the arts), but it is inextricable from its unremitting repetition, its printed vehicle. Home’s book is the terminal point for at least two things: massively distributed global language processing networks and literary language. Its sentences cease to make sense; literally, as Kittler suggests, auguring the end of literature as the practice of the composition of sentences. “Close” up, noise. But, this book can’t only be understood from a distance. If its concern is the work that language does, the “distant” level registers only heuristic filtration processes. Digitality builds a bridge between these levels, providing us with an always-materialised techno-logical route between them. Spam’s incessant fusillades put paid to the distinctions between code and language, old media and new. Is this book inert? Maybe. But considering it as such limits our conceptions of how it works. I would like to suggest that code and software studies could take account of media like this, too, when considering the material networks that form the relays through which computation passes. Rather than looking for media objects, code and software studies might focus on processes of mediation - like digitality - that create specific effects: spamming us with meaning or with meaninglessness.

I don’t agree with Kittler’s alarmist declaration that the birth of language processing machines augurs literature’s death. Claims of the death of things, books included, usually prove themselves premature. Pulled back from this polemical brink, Kittler’s claim suggests something else: that the proliferation of “textual data processing” (Kittler 1986: 159) alters how language is used and how literature is conceived. What is left in the wake of textual data processing’s banalisation? Blood Rites is not a hard and cold piece of literature, as its treatment in this essay might have suggested. It’s crude, base, and often very funny. Language is pliable - iterable - and humour can be fashioned out of anything, spam included. But the meaning of this book fails to cohere completely, disintegrating under the barrages of spam levelled by Home’s compositional techniques. Perhaps the interleaving of humour and technical modes of composition in Blood Rites testifies to humour’s aesthetic-technical ground, given that “like physics, aesthetics is a science whose primary object is signals, the physical materiality of signs” (Siegert 2007: 42). In our digital age - whose birth we might situate with the invention of movable type, if not earlier - our grasp of what makes literature literary will depend on whether our aesthetic categories re-member literature’s materiality and mediality. After butchering his book to extract its technical bones (and to shield its pulp-y body from critical recuperation), Home deserves the last word:

Zero. One. Zero. One. Zero. One. Zero. One. Zero. One. Zero.

One. Zero. One. Zero. One. Zero. One. Zero. One. Zero. One.

Zero. One. Zero. One. Zero. One. Zero. One. Zero. One. Zero.

One. Zero. One. Zero. One. Zero. One. Zero. One. Zero. One.

Zero. One. Zero. One. Zero. One. Zero. One. Zero. One. Zero.

One. Zero. One. Zero. One. Zero. One. Zero. One. Zero. One.

Zero. One. Zero. One. Zero. One. Zero. One. Zero. (Home 2010: 11)

Acknowledgements

I would like to thank the anonymous reviewer(s) of this article for their generous engagements with its earlier iteration; Peter Otto, for his patient and perceptive comments; Tom Sutherland, for his attentive reading; and the editors of this special issue, for their excellent work on the CODE conference and this edition of Scan.

Parts of the research for this essay were conducted using funds from the Amy Gaye Cowper Tennent Memorial Scholarship and from The University of Melbourne’s School of Culture and Communication and Faculty of Arts. I would like to gratefully acknowledge their support.

Biographical Note

Scott Wark is an M.A. student and tutor at The University of Melbourne. His research has been presented at conferences in Melbourne, Sydney and London. His M.A. research uses recent developments in media studies - new notions of networks, new conceptualisations of media, and new developments in the philosophy of technology - to rethink the materiality of the literary medium. He is also a practicing arts writer: his non-academic essays have appeared in catalogues for galleries in Sydney and Melbourne, including Space Oddity (2013) and Innovators 3 (2012) at Linden Centre for Contemporary Arts and Last Words (2010) at 4A: Centre for Contemporary Asian Art.