‘An Object with No Predecessors’? A Computational Reading of J. H. Prynne’s For the Monogram

J. H. Prynne’s 1997 book For the Monogram arguably marks the beginning of what has been termed his ‘late style’: a period of avant-garde poetic output characterised by a shift towards monolithic blocks of text featuring highly disrupted syntax and a vocabulary drawn from a range of increasingly technical and specialist fields. This paper considers whether such ‘high-tech’ writing requires a similarly high-tech approach to reading, describing efforts to interpret the poem using a custom-built computer program linked to the Google Books database. In particular, it examines the theoretical implications of such rudimentary machine reading in light of existing interpretations of the poem by Simon Jarvis and Peter Middleton, focusing on the peculiar aesthetic implications of exponential technological development and arguing that For the Monogram is a text which anticipates and even acts out its own mechanized dissection. Through a complex incorporation of sources ranging from Mother Shipton to computer programming manuals, Prynne anticipates and disrupts any attempt at computational processing, leaving a text which is paradoxically both immune to and deformed by technological progress.

As in later books, the blocks are relatively consistent in width, and they are all untitled, marking the start of a wholesale avoidance of titles-which Prynne refers to derisively in a 1964 interview, echoing his interviewer, as 'handles'-lasting until 2010's Sub Songs. 1 For the Monogram has not received significantly more critical attention than any of Prynne's other books of this period, which is to say it has not received much attention at all. One of the most substantial and sustained attempts to tackle the sequence is Simon Jarvis's 2003 article for the online magazine Jacket, a piece which runs to just 3,500 words and bills itself (over-modestly) as ' a notable Failure '. 2 What is it, then, other than the general obscurity of Prynne's late work, that has prevented scholars from engaging seriously with For the Monogram? A clue might be found in the third poem in the sequence, which seems to have been quoted and cited more times than any other: Select an object with no predecessors. Clip off its roots, reset to zero and remove its arrows. For critics eager to demonstrate the supposedly alien character of Prynne's late poetry, specifically its incorporation of conventionally non-poetic diction, this particular poem is a helpful tool-this is the use to which it is put in Ian Brinton's Contemporary Poetry (2009), a book written largely for students. 4 As Brinton points out, quoting Peter Middleton, the poem makes 'the new communications technology part of [its] field of reference, semantic, visual, and lexical, all broadly at the level of content.' 5 Unfortunately, as in so many readings of Prynne's later poetry, this is where discussion stops. Scholars trained in the profession of English Literature are quick to identify what makes poems such as this unusual and are generally able to make educated guesses about the origins of their language; however, with very few exceptions, they are reluctant to engage with that language on the broad 'level of content' to which Middleton refers. 6 When semantic judgements are made, they are very often of an abstract, second-order character: the deployment of a particular vocabulary is read as a statement, but the actual propositions made using that vocabulary are ignored.
(In many cases, the supposed impenetrability of these baseline propositions is itself interpreted as a second-order statement.) What if 'the new communications technology' could itself help to bridge this gap?
Middleton contends that 'the mobile phone, the internet, the personal computer, and digital image processing have had a greater impact on the way we live than any new set of technologies since the arrival of the motor car, telephone, and radio a century earlier and are genetically modifying the ways we remember and the way we read.' 7 This is no doubt true, but the use of the perfect tense in the first part of this sentence risks concealing the way that these new technologies are themselves developing and superseding each other at an exponential rate. One example will serve to illustrate the pace of change. Middleton Google Books is undoubtedly one of the technologies most responsible for 'genetically modifying [. . .] the way we read' in recent years, with poetry like Prynne's representing prime material for the trial-and-error-based approach to literary interpretation that it facilitates. It is now possible, for instance, to identify a precise source for the sentence 'Select an object with no predecessors' simply by typing it into the search engine, which throws up a 1986 text, Data Structures, Algorithms and Programming Style, by the computer scientist James F. Korsh. 10,11 But what are the limits of this approach? Can it be used to gain a foothold on the elusive 'level of content' referred to earlier, or is it simply a quicker and more efficient way to make guesses about textual provenance? To test this thesis properly, it is necessary to accept the challenge posed by Prynne's adoption of the language of computer programming and actually to write a computer program. Using JavaScript and the Google Books application programming interface (API), I put together a simple script intended, in the discourse of password cracking, to 'brute force' the poem. The script takes plain text as input-in this case, a copy of For the Monogram which had been scanned, run through optical character recognition (OCR) software and tidied up manually. It then strips the text of all punctuation and 'whitespace' and uses it to generate an overlapping list of four-word chunks: words 1, 2, 3 and 4, then words 2, 3, 4 and 5, and so on. (Four words was decided upon as the optimum length, being short enough to cast a wide net but not so short that clear sources would be impossible to identify.) Each chunk is passed to Google Books-if it returns results other than those books in which the poem itself is printed or quoted, their titles are listed and made available for inspection by hand.
One of the most useful functions of this program is to confirm sources that have already been identified or part-identified through conventional reading. In 'The Incommunicable Silhouette', Jarvis puts forward Immanuel Kant's Critique of Pure Reason as an important precursor to the text, using as evidence Kant's frequent use of the term 'monogram' in that work and the phrase 'the scheme of a pure | sensible outline' in the second poem-not a construction used by Kant, but one which Jarvis takes to be a reference to his transcendental schema. 12 Jarvis's argument is As it happens, this very same intertextual link was identified by Middleton in his essay 'Dirigibles', though I had not read it at the time that I wrote the first draft of this article and had never heard Come Out until encountering the apparent reference. 20 What does it mean to have rediscovered this link through the recursive searching of a massive database-a database which, while certainly foreseeable, would presumably have seemed a relatively distant possibility at the time of the poem's composition? 21 Moreover, does the (re)discovery have any interpretative value in the context of a conventional reading of For the Monogram, or is it simply a curiosity to be pored over by specialists? I believe that this sort of automated textual archaeology is, absolutely, of interpretative value, and that the clearest way to see this is through studying the reciprocal effect that new disclosures have on existing theories of source material.
In the case of Poem 7, it is not difficult for a reader to detect references to the body, and particularly to both surgery and industrial meat production: in the course of a repeated 'b-' and 'br-' sound patterning, there are references to bones, blood, brains, bruises and brisket, while the poem also focuses on the processes of cutting and peeling. 22 Adding Come Out to this constellation of concerns is analogous to the insertion of a new explanatory variable in a multiple regression model: under the sign of state violence to vulnerable young bodies, our grasp of the poem's 'meaning' becomes tighter. Nevertheless, it is important to admit that we remain in the orbit of the sort of abstract, second-order reading discussed at the beginning of this article, in which actual propositions are passed over in favour of the broader statements made by the use of certain vocabularies or reference to specific text-external fields.
Before exploring some of the other sources exposed by For the Monogram's computational disassembly, I would like to turn, deliberately, to an ' actual proposition' made by the text-one which seems particularly relevant to the question of how, exactly, the sequence is to be read. There is a regrettable tendency in criticism of Prynne's poetry to search for coded acknowledgements of the work's obscurity, as or 'till'. There is 'lag' between the two, making any attempt to reconcile them a 'broken reach'. If the affronted reader, whatever their ' offence rating', persists with this sort of inadequate strategy, the already 'narrow axis' of their reading will 'taper into counting lucky hits', shaping incidental features of the work into patterns which confirm their 'prior guesswork'. 26 This is the interpretative equivalent of buying on credit, with a ' charge card', a fact of which the lazy but nevertheless 'rueful' reader can hardly be unaware.
If this reading is anything close to accurate-and as long as it remains sufficiently local to avoid being undermined by its own conclusions-then Prynne seems to be making an important statement about the way his late poetry is to be read, or not read. Concluding 'The Incommunicable Silhouette', Jarvis reveals a 'hunch' that the famed difficulty of this work In Jarvis's model, conventional reading strategies, trained precisely to obey 'the functional directives of divided intellectual labour', will clearly be inadequate to the task of reading For the Monogram-they will be exactly that 'broken reach' that cannot but end by 'taper[ing] into counting lucky hits'. But if Prynne's late poetry is not to be abandoned as absolutely unreadable, then some new strategy must be developed which is capable of approaching the work in a movement which mirrors its own 'scrupulous attention' to its 'single quite discrete object'. It is my contention  Under water men shall walk, Shall ride, shall sleep, shall talk.
In the air men shall be seen, In white, in black, in green; Iron in the water shall float, As easily as a wooden boat.

Gold shall be found and shown
In a land that's now not known.
Fire and water shall wonders do, England shall at last admit a foe.
The world to an end shall come, In eighteen hundred and eighty one. 33 For the modern reader, this text obviously bears the one key hallmark of spurious prophecy: verifiability built on vagueness. As such, it might be said to mirror the ' charge card' approach to poetic interpretation laid out earlier-disciples of Mother Shipton, like conspiracy theorists or devotees of Nostradamus, are ' counters of lucky hits' par excellence. Nevertheless, its implicit citation by Prynne 116 years after the supposed apocalypse serves to re-invest it with specific meaning. The third and fourth lines in particular-'Around the world thoughts shall fly | In the twinkling of an eye'-having presumably already predicted the electric telegraph, the telephone and shortwave radio, now seem to refer unequivocally to the internet, and by exten- pamphlet would be beyond the scope of this article; nevertheless, we may note his conclusion, which is that the poem is able to 'pick up and even toy with its own strands'. More fully: The 'meaning' of this poem, or even a meaning or meanings for it, is thus not overtly declared. The method of fluid abstraction prevails over particulate description, the landscape (seascape) and the schedule of colours and visual imagery, drawing these into a domain of conceptualised potential meaning [. . .] which responds to ideation by constant promotion to higher levels of abstraction, so that motions of thought float closely between and within currents of the description but are not specified or captured by them. 36 If this delicate and fine-tuned autonomy is the achievement of Stevens's poem, then its ironic citation in For the Monogram seems to represent a deliberate and demonstrable failure of that achievement. To be sure, this is also a work in which meaning is 'not overtly declared'-but only because there are so many pretenders to the title, pieces of language and other miscellaneous fragments of semiosis thrusting themselves in from all sides. The experience of reading For the Monogram is not, as Stevens has it, 'like being alone in a boat at sea', but more like swimming in a crowded shipping lane-and this is an experience of which the text itself is aware.
To put it bluntly, For the Monogram is a poem written by a human in the knowledge that it will, at some point, be read by a computer. This brings us back to what I believe is meant by Jarvis's ' counter-architectonic working within and against the existing databanks'-not a Wachowskian sci-fi cliché in which The Human Spirit is pitted against the ruthless logic of the machine, but a more ambivalent movement of strategic undermining and sabotage. One of the discourses in For the Monogram that has not yet been noted is that of mathematics, specifically graph theory, as when in Poem 13 we read about ' a non-trivial path from the vertex back | to itself'. 37 In mathematical terms, a graph is a group of objects (vertices) connected to each other in various configurations by a series of theoretical lines (edges); the first real work in the field is considered to be Euler's paper on the Königsberg Bridge Problem, which was in turn spurred by the researches of Leibniz, whose Theodicy provides the epigraph for For the Monogram. Despite this historical pedigree, graph theory was in the late 1990s (and to some extent still is) an exciting field, due primarily to emerging practical applications which had the potential to revolutionize database technology, offering an alternative to the strictly hierarchical rows-andcolumns model that had obtained up to that point. This new horizontal approach might be taken as a metaphor for the structure of For the Monogram, each poem, sentence or word a vertex and the strands of meaning a set of provisional and precarious edges tying them together.
Technological metaphors for what is essentially political resistance are, of course, inherently flawed-however symbolically or formally democratic a new technological development, capital's effective control of research and development ensures that there is always a repressive application waiting in the wings. This is certainly true in the case of graph databases, whose suitability for modelling social relations has led to their adoption by intelligence agencies such as the US National Security Agency (NSA) and GCHQ in their efforts to interpret the vast swathes of metadata collected under programmes such as PRISM. 38 Still, Prynne is not careless enough to remain bound by the restrictions of a single metaphor, and the fate of the graph database does take us close to the heart of For the Monogram's sceptical understanding of technological development and its effect on poetic language. What a computational reading of this sequence reveals is, ultimately, its own insufficiency.
The more data that is collected and the more efficiently the poem's sources are identified, the more obvious it becomes that there is an interpretative gap at the centre of the reading corresponding to Jarvis's 'single quite discrete object, experience or phenomenon'. This is not to say that this object can be grasped through some sort of intuitive leap, nor that computational approaches to reading lack value altogether.
The point is, ultimately, a phenomenological one, having to do with the process of interpretation as much as with the end result. Using a computer's processor and memory to read is never quite the same as using one's own brain, precisely because it implies the outsourcing of the act of processing and its separation from decisions about meaning. 39 This point is derivable from any reading of the poem, even the most scrupulously hygienic, text-internal exercise in 'practical criticism'. Yet it is bound to come across more forcefully in the course of a reading which accepts what For the Monogram tries so strenuously to say about itself: namely, that it is both one thing (a poem) and another (fragments of pre-existing language). From this perspective, the value of a computational interpretation is essentially performative, insofar as its separation of processing from understanding dramatizes a fundamental split which already exists in the text itself, as it does in every text. Facing the poem's hidden centre, Jarvis describes it not only as 'some single quite discrete object', but also, potentially, as an ' experience or phenomenon'. 40 41 In fact, those ' connexions and fissures' would themselves be the objects, rather than serving as phenomenal manifestations of some deeper truth. In this sense, For the Monogram really could be said to 'pick up and toy with its own strands', and the gap between processing and understanding, source text and deployment, would be something like the hinge between finger and thumb which makes it possible.