In the context of a hostile funding environment, universities are increasingly asked to justify their output in narrowly defined economic terms, and this can be difficult in Humanities or Arts faculties where productivity is rarely reducible to a simple financial indicator. This can lead to a number of immediate consequences that I have no need to rehearse here, but can also result in some interesting tensions within the academic community itself. First is that which has become known as the ‘Science Wars’: the increasingly acrimonious exchanges between scientists and scientific academics and cultural critics or theorists about who has the right to describe the world. Much has already been said—and much remains to be said—about this issue, but it is not my intention to discuss it here. Rather, I will look at a second area of contestation: the incorporation of scientific theory into literary or cultural criticism. Much of this work comes from a genuine commitment to interdisciplinarity, and an appreciation of insights that a fresh perspective can bring to a familiar object. However, some can be seen as cynical attempts to lend literary studies the sort of empirical legitimacy of the sciences. In particular, I want to look at a number of critics who have applied information theory to the literary work. In this paper, I will examine several instances of this sort of criticism, and then, through an analysis of a novel by American author Richard Powers, Three Farmers on Their Way to a Dance, show how this sort of criticism merely reduces the meaningful analysis of a complex literary text.

The Allure of Science and Technology

A number of theorists and critics have attempted to incorporate information theory into studies in the humanities: Katherine Hayles (1987), William Paulson (1988; 1991), and Geoffrey Nunberg (1996) have all, to varying degrees, pointed to apparent similarities between the two fields. In describing the hypertext theories of George Landow and others, Nunberg suggests that such theorists talk about ‘information’ rather than ‘content’ because the term ‘incorporates assumptions of [m]obility and transferability in its meaning’ (1996: 107). Roman Jakobson has suggested that there are ‘striking coincidences and convergences between the latest stages of linguistic analysis and the approach to language in the mathematical theory of communication’ (1971: 570), and Paulson maintains that the application of scientific theory to literary texts succeeds because ‘there are features of both model and object that…make the former pertinent to the latter’ (1991: 41). Critics working within the subject of Literature and Science roughly fall into two main approaches to the application of scientific theories in the humanities: the first group claims that there is a deep isomorphism that underlies the two areas; the second suggests that the scientific theory is a powerful metaphor that helps explicate the literary example. I believe however, there is a third alternative which can be usefully mobilised in the context of information theory; one which sees scientific theory, not as having any epistemological connection with literary theory, but as a distinct body of knowledge which can provide unexpected insights into the object of study. Paulson suggests something like this when he asserts that information theory cannot simply be applied directly to the literary text as if it were a new empirical domain, but rather sees its incorporation as the

conceptual use of terms such as foreseeability and redundancy, which have acquired specific meanings in information theory and now enter the field of textual studies with these meanings. In other words a concept such as redundancy that is not inherently limited to the theory of information becomes at once more specific and more complex as a result of its articulation within that theory, and can retain something of this meaning when returned to its original field or imported into yet another territory. This process may in turn lead to new kinds of conceptual thought concerning the literary text… (1988: 65-66).

In this formulation, the connotations which adhere to the specialised terminology of the scientific theory when it is transported into the literary theory can reveal assumptions and misapprehensions that were not previously obvious. Writing in 1960 amid the first rush of interdisciplinary enthusiasm for information theory, Jakobson describes such a productive dialectic in relation to the term ‘redundancy’ as it is used in both information theory and linguistics. In information theory, redundancy lost its general connotation of excess or superfluity, and stood instead for the repetition within a message that ensured clarity, even in the presence of noise or impediments. Thus, redundancy was not waste, but an essential and integral part of the message. Jakobson points to a tendency among some theoreticians to ignore the distinction between redundant and distinctive features in phonetics, and claims that the re-introduction of the term is correcting this misconception: ‘[t]he prejudice treating the redundant features as irrelevant and distinctive features as the only relevant ones is vanishing from linguistics, and it is…communication theory…which helps linguists to overcome their biased attitude’ (1971: 572). I will examine the congruence between information theory and literature from this perspective. Any attempt to reduce a literary text to the mere communication of a message (although it can be that as well) ignores most of what it is that makes a literary text worth studying. Thus, it is not my intention to analyse the text as a ‘message’ that can be transmitted or misinterpreted, but to examine terms such as ‘redundancy,’ ‘equivocation,’ and ‘noise’ in order to see what additional insights they can produce owing to their incorporation in information theory.

An important place to commence this investigation is with a definition of ‘information.’ This seemingly unproblematic term is mobilised in a number of different discourses, in different ways, and with very different effects. The theorist who tries to incorporate information theoretical definitions carelessly can find the term breaking down in unexpected ways. The word ‘information’ derives from the Latin, and, while it is a substantive noun, its earliest usages in English (around the fourteenth century) were synonymous with the continuing aspect (or ‘present participle’) of the verb—‘informing’—and defined it as the action of moulding the mind, or imparting knowledge. It was not until the mid-fifteenth century that the word began to refer to the actual knowledge that was communicated. In time, it also gained specific definitions within legal discourse, and the sense of ‘intelligence’ pertaining to spying activities. In 1937, specifically in an information theoretical context, the word was first used in a sense that separated it from the person informing or being informed; thus, it came to be seen as a discrete object that could be transmitted. Increasingly, information is seen as an entity (possibly immaterial) that exists objectively. In a few hundred years, ‘information’ moved from being an active agent of change to being a passive object subject to mechanical transmission and modification. Geoffrey Nunberg suggests that there is a subtle distinction between two types of information: ‘particularist’ and ‘abstract.’ The particularist sense is the definition we find in the Oxford English Dictionary, but beyond this is a sense that refers to ‘a kind of intentional substance that is present in the world, a sense that is no longer closely connected to the use of the verb “inform,” anchored in particular speech acts’ (1996: 110). He maintains that it is this sense of the word that is mobilised in discussions of the new technologies.

This abstract sense of information gained its impetus in the early1940s when Claude Shannon defined information, not as an active agent, or even as an existent element to be transmitted, but as a mathematical formula based on the probability of a signal arriving at its destination intact. In terms of information theory, Shannon had no interest in the meaning of the message (Shannon & Weaver 1962: 3), or whether there was any significance in the content, and, in this sense, his theory is misnamed. Yet, despite the complete absence of any theorisation of ‘meaning’ within the specific usage of the term, Shannon and Weaver’s ideas soon gained a centrality within the growing discipline of communication studies. It is in this productive period that the Jakobson article mentioned earlier was written. Despite the similarity of terminology, and the communication theorists’ fascination with Shannon and Weaver’s model, there are very few conceptual links between communication theory in the humanities and information theory in the sciences. Where Shannon elided references to meaning completely, the fundamental purpose of communication theory is to explicate the transmission and production of meaning. Yet, this essential difference has not deterred theorists such as Hayles and Paulson from unproblematically incorporating Shannon’s model into general theories of literature or culture.

For the communication theorists, as well as later critics, the most valuable part of Shannon and Weaver’s theory was the concept of noise. Quite simply, noise is any external input into the channel between sender and receiver, which has the potential to obscure the signal. One of the main aims of information theory was to find ways to overcome noise: to increase the probability of the message arriving in a form that could be accurately decoded. Paulson suggests that ‘literature is a noisy transmission channel’ (1988: ix), and thus incorporates many of the aspects of information theory into his analysis. One of these elements is the possibility of meaning deriving from randomness: ‘[u]nder the right circumstances, noise—from whatever source—can create complexity, can augment the total information of a system’ (Paulson 1988: 73). In its usual terms, however, an increase in noise results in a decrease in determinable meaning. A major goal of information theory has been to reduce the occurrence of random input and to allow for the replacement of lost data. The element isolated by Shannon to overcome the problem of noise was redundancy. This can range from simple repetition to complex formulae for error checking and data replacement. In natural languages, the degree of redundancy is high, as evidenced by the ability of most people to project accurately the meaning of an incomplete sentence or to replace missing words in a text. The important point is that while information and redundancy may appear to be mutually exclusive, they are both intrinsic and mutually dependent parts of a text:

Information cannot inform in a system devoid of redundancy, and redundancy remains the pointless repetition of the same in a system devoid of variety. The kind of organization that emerges from disorder, and that can be contrasted in significant ways with disorder, cannot be described by a single concept from information theory, such as redundancy or information, but must contain within itself structures that exhibit both redundant order and informational complexity. (Paulson 1988: 73)

This pair of dichotomies—signal/noise and information/redundancy—forms a crux at the centre of the criticism that tries to apply information theory to the literary or cultural text. The tension between these two axes creates a space in which the informational, literary object can be analysed. Paulson, for instance, suggests that literary language is a ‘vehicle for the creation of new information,’ and for this to be true:

literature must be, to a degree, both obscure and repetitive; were it perfectly clear, nothing new could arise in the transaction between author and reader, and were each of its utterances original, its relative obscurity would destroy information faster than it could create it. (1988: 101)

This is, of course, a truism. However, Paulson expresses his literary theory in terms of the mathematical function of redundancy and signal, as if the intelligibility of the text is the only issue. A fundamental point which both Paulson and Hayles take from information theory is the problematic relationship between information and redundancy, between novelty and certainty: ‘[m]aximum information is conveyed when there is a mixture of order and surprise, when the message is partly anticipated and partly surprising’ (Hayles 1990: 53); ‘a received message in effect performs a selective operation among possible states of awareness or readiness or action in the receiver; richness of meaning corresponds to the complexity of the selective operation’ (Paulson 1988: 61). The first point to make is that these statements are true whether or not information theory is part of the equation, and the application of the mathematical theory merely obscures the point. More problematically, however, this approach conflates textual complexity with semantic complexity, and it does not necessarily follow that a complex idea cannot be expressed in a simple form. The problem with this approach is that the redundancy that Shannon discusses has very little to do with the surprise with which a text may confront readers or their ability to anticipate and select the appropriate response. Ultimately, therefore, these approaches are sterile: they cannot tell us anything more about the text than what we can learn from more conventional literary theories.

This type of terminological conflation is nowhere more apparent than in discussions surrounding the troublesome word ‘equivocation.’ In his mathematical formulation, Shannon used the word to describe noise. Perhaps this should not surprise us: he had earlier conflated the terms ‘information’ and ‘entropy,’ and was notorious for his highly idiosyncratic use of terminology. Yet Hayles and Serres have both pointed to the similarity of terminology in order to conflate the literary meaning of the term, with its specialised usage in information theory (see Hayles 1987; 1988; and Serres 1982a, especially Chapter Seven). Shannon’s choice of the word ‘equivocation’ suggests that he was concerned with the relative ambiguity of the signal itself, with the link between accidental interference and the (theoretical) purity of the original message. In information theory, equivocation is a result of data that is infected by noise between the sender and the receiver. This is a very different thing to equivocation in literature which results from the intrinsic nature of language—not something added after or during transmission, but an element of language that is not only inherent in the semiotic process, but actually constitutive of that process. As Serres’s formulation expresses it, equivocation ‘result[s] from noise’ (1982a: 77).

There are interventions, however, which use information theory in interesting and productive ways. Serres, for instance, provides an interesting perspective on noise. He draws on the conflation of sense that sees the French word ‘parasite’ carrying not only the biological meaning, and the sense that in English is carried metaphorically—the parasite as a human scrounger—but also the sense of the interruption of a transmission or signal, which is translated in English as ‘static.’ In this way, he draws connections between biology, human relations, and information theory to produce a social analysis. His thesis is that every relationship is parasited in one way or another: the tax-collector parasites the farmer’s wealth; rats consume the tax-collector’s left-overs; guests parasite the host by paying for their lodgings with mere words; a noise disrupts the system. ‘There is no system without parasites. This constant is a law’ (1982b: 12). Yet, according to Serres, this law is also a circularity. Considering Aesop’s parable of the city rat and the country rat, he draws a diagram of the levels of parasitism: the farmer is parasited by the city rat, who is parasited by the country rat, his guest, and in the end the rats are scared off by a noise (parasite = static) at the door, a noise caused, in all likelihood, by the farmer (1982b: 3-14, and passim). So the farmer, the first host in the chain, becomes the parasite of the parasites. Serres points out that the influence that has changed this system from a simple linear one to a complex, circular arrangement is the input of noise:

noise gives rise to a new system, an order that is more complex than the simple chain. The parasite interrupts at first glance, consolidates when you look again…. The town makes the noise, but the noise makes the town (1982b: 14).

As a general principle of organisation, he then points out that the system of interacting parasites thus produced is much like any ordinary system, subject to its own interruptions and breakdowns, and that the ‘parasite is the essence of relation. It is necessary for the relation and ineluctable by the overturning of the force that tries to exclude it’ (1982b: 79). Ultimately, the role of the parasite within the system is ambiguous: ‘[w]ho will ever know if parasitism is an obstacle to its proper functioning or if it is its very dynamics?’ (1982b: 27). Serres shows how equivocation in information theory can be seen as something inherent, rather than added, and in this sense, the two distinct meanings of the term can be seen to mutually enrich the concept. Equivocation (in the literary sense, but being mindfulof the added connotations from information theory) is something inherent in every text, but it becomes more obvious and has more significant effects with the increasing complexity of the text. It is due to the slippage of signifiers, and produces the richness and complexity that some would say is the defining feature of a literary text. Equivocation as noise is a concept that is not very different to this formulation. Noise informs a text, sometimes to its detriment, but more often in productive ways.

These two interventions into information theory demonstrate the sort of fascination that some literary theorists have for science and the way technology is supposed to advance the role of the critic. In the second part of this paper, I will examine a novel that seems ripe for the kind of analysis that Hayles, Paulson, and others have performed in the terms of information theory. Three Farmers on Their Way to a Dance is a novel that deals with perplexing issues in a complicated and convoluted way. As I will show, the complexity of the style can alienate the reader, and in this sense can be considered ‘noise’ that impedes the understanding of the novel. But as I have said, a direct application of information theory to a literary text is ultimately sterile. The conflation of literary and scientific theories inevitably produces a limited and totalising object that is exceeded in many ways by the literary text it was meant to analyse. Three Farmers exemplifies what I have chosen to describe as the ‘heterotemporal’ novel, a text where the collision of temporalities produces a signifying space that is outside the generalising capacity of literary theory.

The Richard Powers Case

Powers’ novel displays a convoluted form that is distinctive of his works. It consists of three intertwined and problematically connected stories that revolve around a single, somewhat mysterious photograph of the eponymous farmers, taken in 1914, a few months before the outbreak of the First World War. The first story is an historical narrative that concerns the farmers, their adventures during the war (two of them are killed), and the events and people associated with the era. The second story is a sequel to the first. It is set in 1984, in Boston, and concerns a young man, Peter Mays, a descendant of one of the farmers, who discovers a personal legacy encoded in the photograph. The third story, told by an unnamed, first-person narrator, is an extradiegetic [1] narration in which an author—a thinly disguised Richard Powers—decides to write the novel I have just described. This first-person narrative is interspersed with essays on twentieth century history; the theory of photography and mechanical reproduction; biographies of Henry Ford and Sarah Bernhardt; and other non-fictional elements that were presumably relevant to the researching of the novel. Thus the novel contains two aspects, the ‘essayistic’ and the ‘imagined,’ which together bring the photograph into three-dimensionality (Powers 1985: 334). The themes of the novel revolve around the First World War and the calamitous end of an era. Both the extradiegetic narrator and Peter Mays become obsessed: the narrator with the photo; Mays with a sense of anachronism (first relating to a mysterious redhead, then to an impossible photograph of himself with Henry Ford). Behind these obsessions is a drive to understand some indefinable essence the photographs seem to carry, a meaning that is interpretable to the close reader. Powers’ narrator sees the web of discourses that surround the photograph as a way ‘to finally put to rest, after seventy years, the destructive residue of the Great War’ (1985: 334).

The structure of the novel exemplifies some of the problematics of heterotemporality. The two third-person narratives are historically continuous, but temporally remote. The effects in the latter narrative are traceable to causes in the first, yet the action of one takes place in 1914-1915; the other in 1984, and there are no characters common to both. Further, the actions and discussions in the extradiegetic narrative at once parallel and problematise the other two by introducing characters with structural similarities to those in the other narratives, who nevertheless perform very different functions in the novel. These disparate threads are offered in alternation—each of the three story lines cycles in a predictable order: Chapters One, Four, Seven, etc. follow the story of the extradiegetic narrator; chapters Two, Five, Eight, etc., the three farmers, and so on. Selectively reading every third chapter produces a coherent narrative that largely stands on its own. The relationship between the three threads in the novel is a tension that holds the narratives in position. Each is completely explicable on its own, yet the additional meaning provided by the parallel threads at once enriches and problematises the meaning of the threads. None of the threads in the novel can be considered as more important than any other, whether contemporary or historical, extradiegetic or essayistic, and while the events in the contemporary thread are subsequent to those in the historical thread, their importance belies the subordination implied by the term ‘subsequent.’ It is this idea that I call temporal flattening. While the fabula of one thread is located some seventy years in the past, its sjuzet is indistinguishable from the contemporary threads, in either style or diction.

This flattening of temporality is perhaps seen nowhere more clearly than in Powers’ use of tense. Most of the novel is narrated in the past tense, whether it is first- or third-person narration, narrative or essayistic. The only exceptions are Chapters Two and Twenty-six, the first and last chapters in the ‘farmers thread,’ which tell and re-tell the story of the taking of the original photograph and which are narrated in the present tense. The effect of a present tense narration is to imply the ‘presence’ of the narrator to the events being narrated by conflating the temporal and spatial immediacy inherent in the term ‘present.’ In narratological terms, the order, duration and pacing of these two chapters are no different from any others in the ‘farmers thread,’ yet the use of tense suggests a different temporality. The episode of the photographer introduces the story of the three men, yet it also stands alone, atemporally separated from the rest of the narrative. It creates a paradoxical sense of unreality that is at odds with the materiality of the photograph that provides the only physical link between all three story lines. This unreality is accentuated by several passages where descriptions of the events and descriptions of the photograph meld into one:

[t]hree men, only one of them native to the area, are photographed as if they are walking to a not-quite fair. One lags a little behind, cane at a rakish angle, undisciplined hair curling out from under a beaten-up hat. His lip shows the first sign of down. His jaw seems to contain an extra joint just above the chin. (1985: 28)

The action of the three farmers walking to their dance is now indistinguishable from the photograph that captured the image. In a sense, the historical referent of the action is elided by its signifier. An historical aura still exists in the mechanical reality of the photograph—it is in monochrome; the shading contrast is not well developed; and it shows marks and imperfections that are consistent with its age—but the historicity of the action is now only apprehensible through the photograph, just as the historical events narrated in the novel can only be approached through the temporal frame of the present. This is not only in the most obvious case of the three intertwined narratives and the interpolated essays, but within the chapters as well. In Chapter Two, the narrator describes the farmers as they approach the awaited dance, and begin to hear the music in the distance:

[s]ound travels at about 1089 feet per second at freezing, and increases about 1.11 feet per second for each degree Fahrenheit rise in temperature. This crisp May Day is a perfect 68 degrees, so the speed of sound today is around 1129 feet per second. In five seconds, the music—whether it is Bavarian Beer Hall or Viennese Woods—travels 5645 feet, the little over a mile that the three boys are from the site of the May Fair (1985: 20).

This editorial interpolation creates a kind of disjunction in the text: it points to the gap between the narration and the events by describing historical events in modern, scientific terms. It also contributes to the effect of information in the novel, which seeks to overwhelm the reader with irrelevant or barely relevant data. But perhaps more importantly, it becomes a metaphor for Powers’ view of time and history:

[m]odulation replaces modulation; in two measures’ time the song the boys hear is no longer the song being made. The boys have only outdated music with which to create a belated present…. The realities of the past become true only when they intersect the present. Then, only, do they become present, known, regardless of what has happened to them since (1985: 21).

Here, Powers suggests that the cataclysmic events of the ‘farmers thread’—the deaths of two of the farmers, metonymically standing for the eight and a half million people killed in the First World War—can only be made meaningful through their connection with contemporary events. Kimberley Greene, an actor in the ‘contemporary thread’ who impersonates famous historical figures, personifies this process in the way she delights in the temporal connections her rather unusual existence affords. When Peter Mays appears backstage after a show, she immediately recognises his likeness to an ageing newspaper clipping in her historical collection: ‘Thanks for coming backstage; when I first looked up at you, I had a lovely jolt of synchroneity’ (1985: 235). In response to the puzzled look Mays gives her, she adds, ‘all times at once’ (1985: 235), which is a good description of one of the aspects of heterotemporality. Together with these sorts of interpolations, the braided narratives and the extradiegetic essays combine to demonstrate this temporal flattening. With all signifiers competing for attention in a noisy pastiche of reality, it is inevitable that some of the information will be obscured. As Powers’ extradiegetic narrator says:

I had lost both the mystical impression and the physical sense of that Detroit photo. I had come across those faces by accident, but it was also the gradual accumulation of daily accidents that blurred and overexposed those faces in my mind (1985: 77).

This introduces the theme of accident or chance. The noise that obscures the signal arrives by chance; it was accidentally incorporated into the message out of the chaos of random background noise. However, as Powers’ narrator indicates, this random noise need not be meaningless: the discovery of the photograph while he was looking for something else turned out to be something centrally important. This productive aspect of noise in the text is quite unlike the sterile definition provided by information theory. Here it stands for that indefinable human capacity to find unexpected but meaningful correlations between disparate items rather than the reductive, mathematical ratio of data to information. Noise can add to the complexity of a message at the same time as it obscures and problematises the original signal: the stray gust of wind that blew mould spores over Alexander Fleming’s bacteriological experiment was ‘noise’ that ruined the integrity of the experiment, but, without it, penicillin would not have been discovered.

I have been emphasising the principle that the introduction of random noise into the system causes the message to spin out of the author’s or readers’ control, producing effects that are built into the structure of language itself. As Jeremy Campbell suggests, the ‘proper metaphor for the life process may not be a pair of rolling dice or a spinning roulette wheel, but the sentences of a language, conveying information that is partly predictable and partly unpredictable’ (1982: 12). Jacques Derrida claims that there is a fundamental undecidability inherent in situations where relations of force create a controlling context—be that, for example, discursive, ethical, or political. The meaning of a text, say, or the possibility of action in a political situation always takes the form of “a determinate oscillation between possibilities” which are always pragmatically determined (Derrida 1988: 148). This undecidability is the basis of meaning:

[w]e certainly count on the calculating capacity of language, with its code and game, with what regulates its play and plays with its regulations. We count on that which is destined to random chance (ce qui destineau hasard), while at the same time reducing chance (Derrida 1984: 4).

Three Farmers is open to a great deal of chance because of its complexity. Minor events, or seemingly irrelevant comments in the early chapters, escalate into important elements, the early clues by then lost in the plethora of detail. It is this complexity of the signal, combined with a faltering human memory that leads to the increase in the text’s entropy. In Three Farmers, the Peace Ship story is told from three different perspectives (1985: 119-127; 301-303; 311-14), each one not only adding to the story, but also problematising it through the addition of details. A privately minted commemorative penny bearing the likeness of Henry Ford, and the words ‘Help the Other Fellow’ in place of ‘In God We Trust’ is described in a number of places (1985: 119; 210; 329-330) by different narrators and with different effects in the story. According to information theory, the greater the degree of redundancy in a text, the lower the amount of information, yet in this novel it has the opposite effect. Events retold some hundreds of pages after they were first related have changed in significance and meaning because of the developing story line. The Fordian pennies, when first mentioned, are an historical curiosity of limited relevance, yet they become in the final chapters a device on which the plot turns. This proliferation of fine detail adds to the general richness of the text, but is usually only discovered explicitly on re-reading, when the early clues are recognised for the significance they contain.

This can be explained quite easily in terms of a generalising literary theory such as narratology. Gérard Genette even manages to incorporate achronic structures into his theory of narrative order as a special type of anachrony that is not anchored in either the past or the future, but connected only to the accompanying commentary (1980: 83). Yet the heterotemporality of Three Farmers contains an achrony that is properly irreconcilable with all other aspects of the novel. I have described the structure of the novel as being a story with a sequel, accompanied by an extradiegetic narration. One of the farmers, Peter Schreck, in order to escape military service, masquerades as a journalist and in this capacity attends a press conference given by Henry Ford during the Peace Ship debacle. Ford, stung by the ridicule of the other journalists, and finding an ally in the farmer, who is really only interested in machinery, sets up a trust fund for Peter, redeemable by a direct descendant who is in the same profession. The story follows the adventures of Peter’s illegitimate son, who marries, and in the 1960s migrates to the United States, eventually producing a daughter, who in turn has a son—Peter. This person is recognisable as Peter Mays from the ‘sequel’ and that plot line largely concerns Peter’s becoming aware of and claiming the Ford bequest. The two narrative threads can be seen as what Genette calls ‘proleptic analepses’ and ‘analeptic prolepses’ (1980: 83). In spite of their concurrency, one thread anticipates the other, which, in turn, gains much of its sense retrospectively from the first. However, as I have said, there is a narrative interpolation that makes impossible this neat characterisation. The son of Peter Schreck, named Peter Hubertus, is jailed during the Second World War, and passes the time by vividly inventing the geography and inhabitants of a fictional city based on the photograph of the three farmers that his mother had given him. This invention remains a habit after the war:

[h]is daughter and her husband had no children, and so the line ran out in the New World. Peter Hubertus did not like this end to the story, so in the Catskill evenings he devised another. He had brought his mother, wife, and young daughter to America just after the war—or better, just before. They escaped in time. They stood before the immigration officer at an Ellis Island of an earlier day, one he had read about in texts. The officer demanded his name. Hesitant, he spilled out:

—Peter Hubertus Kinder Schreck Langerson van Maastricht.

The immigration officer pressed his temples and responded curtly:

—Peter Mays (1985: 321).

The revelation that Peter Hubertus’s daughter (ostensibly Peter Mays’s mother) had no children introduces a disjunction in the text that is irreconcilable to the neat arrangement of story and sequel. The two third-person narratives—stories written by ‘Richard Powers,’ the extradiegetic narrator—cannot exist in the same conceptual space. The logical inconsistency of Peter Mays’s simultaneous existence and non-existence cannot be rationalised as the old man’s imaginings—the account holds too many anachronisms. Neither does the text allow for the three farmers to be in the imagination of Mays, whose lack of knowledge of his family’s history provides the central puzzle to his quest for the Ford inheritance. The only possible response is that the passage is an irreconcilable disjunction in the text: as it cannot be rationalised out, it must stand. This is a case which is comparable to the novels of Alain Robbe-Grillet, another exemplar of the heterotemporal novel, whose temporal ordering Genette admits cannot be reconstituted because their ‘temporal reference is deliberately sabotaged’ (1980: 35). This disjunction, in spite of its transgression of the principles of temporal ordering, does not make the narratives meaningless. On the contrary, the problematising of meaning adds to the production of meaning through the slippage of signifiers.

Much literary criticism has been informed by the assumption that the duty of the critic is to eliminate the noise in a text. The work of de Saussure, Derrida, and others, has created the possibility for issues surrounding the concepts of noise and chaos to be incorporated into theories of meaning, yet there are very few theorists who seriously engage with this question. Three Farmers on Their Way to a Dance sits in that conceptual space between the over-determining redundancy of the text, and its proliferating complexity. This space is held open by the capacity of language to convey multiple meanings and to provide a deep-seated indeterminancy at the level of signification.



[1] For useful explanatory discussions of this and subsequent technical expressions, see Gérard Genette (1980) and Manfred Jahn (2005)



Jeremy Campbell (1982). Grammatical Man: Information, Entropy, Language, and Life (London: Allen Lane)

Jacques Derrida (1984). ‘My Chances/Mes Chances: A Rendezvous with Some Epicurean Stereophonies,’ in Taking Chances: Derrida, Psychoanalysis and Literature, ed. J.H. Smith & William Kerrigan (Baltimore: Johns Hopkins University Press), pp. 1-32

Jacques Derrida (1988). Limited Inc, ed. Gerald Graff (Evanston: Northwestern University Press)

Gérard Genette (1980). Narrative Discourse: An Essay in Method, tr. J.E. Lewin (Ithaca: Cornell University Press)

N.K. Hayles (1987). ‘Text Out of Context: Situating Postmodernism within an Information Society,’ Discourse, Vol. 9, pp. 24-36

N.K. Hayles (1988). ‘Two Voices, One Channel: Equivocation in Michel Serres,’ SubStance, Vol. 57, pp. 3-12

N.K. Hayles (1990). Chaos Bound: Orderly Disorder in Contemporary Literature and Science (Ithaca: Cornell University Press)

Manfred Jahn (2005). Narratology: A Guide to the Theory of Narrative

Roman Jakobson, Roman (1971). ‘Linguistics and Communication Theory,’ in Selected Writings, ed. Stephen Rudy (The Hague: Mouton), Vol. 2, pp. 570-579

Geoffrey Nunberg (1996). ‘Farewell to the Information Age,’ in Future of the Book. (Berkeley: University of California Press), pp.103-138

W.R. Paulson (1988). The Noise of Culture: Literary Texts in a World of Information. (Ithaca: Cornell University Press)

W.R. Paulson (1991). ‘Literature, Complexity, Interdisciplinarity,’ in Chaos and Order: Complex Dynamics in Literature and Science, ed. N.K. Hayles (Chicago: University of Chicago Press), pp. 37-53

Richard Powers (1985). Three Farmers on Their Way to a Dance (New York: Beech Tree & Morrow)

Michel Serres (1982a). Hermes: Literature, Science, Philosophy, ed. J.V. Harari & D.F. Bell (Baltimore: Johns Hopkins University Press)

Michel Serres (1982b). The Parasite, tr. L.R. Schehr (Baltimore: Johns Hopkins University Press)

C.E. Shannon & Warren Weaver (1962). The Mathematical Theory of Communication (Urbana: University of Illinois Press)