Speculation and the End of Fiction

1 Data

Boldly summarising Elena Esposito’s ground-breaking work on the history of probability calculation (Esposito 2007), one could state that probability strictly speaking does not exist. We might thus well say about a future event that the expectancy of its occurrence is 75%. But after the fact, when said event has occurred, or not occurred, the correct number will be 100%, or 0%, respectively. Probability, in other words, is not about something that actually exists, e.g. the event that has occurred or not occurred in the future, but about a certain mode of non-existence, the future as it has not yet happened. Probability does not say anything about “the future’s present” (100 or 0%), it solely deals with the otherwise elusive matter of “the present’s future”, where such a thing like a three-quarter fact can exist. In the terms of Luhmannian systems theory, which is where Esposito takes her point of departure, probability is a “second-order system” and thus situated at an ontological remove from the fullness of a first-order system. In the case of probability, it lacks, unsurprisingly, the fullness of data pertaining to what will eventually be the case in the future’s own present.

If speculation is on the rise, however, it becomes visible in the way such second-order practices insinuate themselves in the making of the future. Recently, the French newspaper Le Monde ran a series of articles on “predictive medicine,“ insistently pushing the new epistemological conundrum in biomedicine: should we cure an illness before it has a symptom, if it can speculatively be predicted to occur with, say, 75% certainty?^[1] In such a case, a pre-emptive therapy would effectively foreclose the possibility of any (let alone firm) evidence; future facts – like symptoms being there or not, 100% or 0% – would themselves have become a kind of second-order phenomena, depending entirely on the way we would already have acted in the present’s future.

This is a matter of data. We obviously don’t have any first-order data about the future’s present. But we have more data than ever deriving from the past and the present. And stipulating a relative historical stability of the logic of living bodies, it does indeed seem tempting to exercise medical intervention in the somehow spectral dimension of futurity. In most cases, such clinical decisions will rely heavily on big numbers: so and so many individual cases of whom we know this and that. This is about significant numbers, and about an adequate mining of what kind of information will be relevant. Errors, of course, might occur; it might be that we are mining the wrong data, like when we are misled by overwhelming correlations in a data set without inferring the right causalities behind them, and it might be that we run up against future contingencies, perhaps a simple and unobtrusive cure henceforth available at the time when the symptom appears.

In terms of errors, in other words, we are dealing with “known unknowns” of correlations and “unknown unknowns” of contingencies, in Donald Rumsfeld’s strangely perspicacious words. But in both instances, the amount of data makes a difference. In the first, reliability increases enormously once it goes towards “all”. It has become a common saying that it is far better to have many data than having good data; a few years back Google boasted to be superior at detecting flu pandemics based on big numbers of (perhaps) flu-related searches on the net to the traditional surveillance mechanisms of the health authorities – a position that has afterwards been contested, albeit not denying the power of gathering big amounts of data, their provenance and validity in the individual case notwithstanding.^[2] Similarly, the building of large databases has proven vital also to other kinds of medical treatment. It is remarkable, for instance, that we have seen such a radical increase in survival rates for cancer patients in recent years, which has taken place without any real breakthroughs in the understanding of the disease or in the therapies, still heavily dependent on the gross cell poison methods. The progress seems to have come about mostly due to a meticulous mining of the records on how the different chemotherapies work, allowing for a still more precise administering of a still quite dumb measure, to the degree that today statistics is an integral part of the therapy, and every given therapy is an integral element of the statistics. Big numbers are simply very efficient, “instead of searching for the causes […], we now seem much more interested in correlations,” says Jacques Attali, and adds: “If diagnostics remains linked to the progress of medical knowledge, prognostics is not merely a result of a better understanding of the causes underpinning the different pathologies, but also the consideration of statistical correlations in big data sets of previous cases” (Attali 2015, 121). Speculation thrives on big numbers as never before, and incredible things are accomplished through the mining and algorithmic modelling of correlational patterns, even though our understanding of the law of big numbers has probably not progressed much since. The man without qualities attempted to explain it to his friend Gerda:

This is somehow vaguely known as the law of large numbers. Meaning more or less that one person commits suicide for this reason and another for that reason, but that when you have a very large number the accidental and personal element cease to be of interest, and what is left is – well, what is left? This is what I want to ask you. For, as you see, what remains is what every one of us, as a lay person, without more ado refers to as the average, the thing, that is, as to which nobody has the slightest idea what it really is. Allow me to add that attempts have been made to explain this law of large numbers by logical and formal means, so to speak, as a self-explanatory fact. On the other hand, it has also been asserted that such regularity of phenomena that are not causally related to each other cannot be explained at all by the usual intellectual operations, and, besides many other analyses of this phenomenon, the assertion has been made that this is not only a matter of single events but also of unknown laws of totality. I don’t want to bother you with the details, which I no longer have at my fingertips anyway, but without doubt it would be very important to me personally to know whether what is behind this is laws of collectivity or whether it is simply that by some irony of nature particularity arises out of the fact that nothing in particular happens and the highest meaning turns out to be something that can be got at by taking the average of what is most profoundly senseless.

(Musil 1955, 228f.)

Moving from the question of correlations to that of contingency, another feature of big data comes to show which is less one of understanding than of attention. Sub-merged by the avalanche of data and metadata conjured up in the big wave of digitisation, our present is better documented than ever before. And based on this thick present, we can indeed generate predictions about virtually everything. This habit of mining contemporary and past data in order to predict future events does, however, invite us to see the future’s present as a reduplication of the present’s future. Reinhart Koselleck famously distinguished between “the space of our experience” and “the horizon of our expectations” in order to gauge the understanding of the rhythm of historicity in different contexts, the ways in which memory and hope interact, in the writing of history as well as in the making of history (Koselleck 1988, 79). In a world of wall-to-wall prognostics based on algorithmic processing of huge databases, we are now reaching a state where, when glancing at the horizon of what we might expect, we see nothing but the well-documented space of experiences already done. Prediction technologies have made us awfully good at forecasting, at looking into things to come, but it is as if we no longer look at a wide horizon, but only into a narrow zone where what we know is prolonged, a future with a narrow scope and a high resolution, as it were. Feeding on such predictions, our historical imagination itself might eventually suffer, and with it our very capacity to aspire – the demise of utopia in an age of algorithms.

It is one thing, however, to loose the attention towards contingency, but another to actually eradicate it. The crucial point is when we have set up a scenario of a future to come, a probabilistic modelling of what is likely to ensue from the present, and then actually react to it. This is, as Brian Massumi has put it, “a futurity with a ‚virtual power to affect the present quasi causally” (Massumi 2005, 35). This reverse order of things where the future – as established by intelligent machines – acts on the present is one of heightened foresight, but it is also an order of what Attali calls “predictive tyranny”, where “the future functions as a dictatorship imposing agency in the present” (2015, 145).

At this point, prediction morphs into a new guise of fatality, i.e. when the present is first projected onto the future and the future then projected back onto the present. Predictive medicine could again be invoked as a case in point, or predictive policing,^[3] often dubbed “precrime” with reference to Steven Spielberg’s visionary Minority Report from 2002, targeting the disease you will have developed, or the crime you will have committed (a mode of reasoning, as it appears, that we haven’t yet completely been able to adapt to the linguistic tense system). The futurity of algorithmic data modelling seriously tinkers with our sense of time, very conspicuously in the areas of medicine and policing, but also in the mundane ways in which big data processing is seeping into our everyday practices. We generate, store and process piles of information regarding our whereabouts, our communication, our browsing and buying, our information requests – or rather, the providers of electronic services generate, store and process these data, transposing our lives into privately owned profiles.

We have “digital doubles” that not only track us, but also address us. At a certain point, when we are well enough documented, the algorithmic morphing of our personae will be able to produce our futures, when incrementally our range of agency, linked up to the digital persona as we are, is getting defined by the algorithms: if you liked this, you might also like that, you should meet this person, you should avoid gluten, unfortunately you cannot be insured, you cannot be granted a visa, etc. In this sense, archives don’t neutrally record content, they also produce a range of possibilities in the present. The tinkering with time eventually becomes a tinkering with subjectivity, directing us towards specific futures that emanate from our algorithmically concocted profiles and prescribing the future selves we don’t even know of yet. Or put even more sternly by Bernard Stiegler: “When the subject is affected in advance by its double, it renders reflexive subjectivity as we know it obsolete, the ‘subject’ always comes too late, and it never has to account by itself for what it is or what it could become” (Stiegler 2015, 201). At this point, we are no longer just having digital shadows, but the other way around: we become shadows of digital personae, like in Hans Christian Andersen’s romantic tale about the shadow progressively taking over the life from its former master.

And based on what? On correlations in big data sets. And therefore, more importantly: based on algorithms identifying and weighing what is to be correlated to what. The medium of speculation here is the algorithm, illegible to most of us, revealing a profound and alarming new illiteracy – were it not that most of them are privately owned anyhow, in which case it really doesn’t matter whether we can read them. So what happens to probability in the age of big data is a systematic short-circuiting between the future’s present and present’s future where the latter is invested in the production of the former with hitherto unprecedented systematic means. The ubiquity of information harvesting and the industrialised processing of it, as it now imposes itself on decision making over a broad range of societal sectors, together seem to be about to change not only probability, but even futurity, by remediating it and shifting it from a speculative reasoning that can be tested on the evidence of a situation to come, into a veritable speculative mode of production, producing futures that align with the correlations written into the new languages of algorithmic modelling.

2 Money

Paper money is probably the most obvious medium of speculation we can think of: a token of value assured by the state, facilitating the circulation of investment capital, and hedged, as it were, by a guarantee that it retains its value. As commonly remarked by sociologists of money well versed in Marx, from Simmel to Sohn-Rethel, these two basic properties of money – that it represents a value, and that it guarantees this value – identify two distinct axes of operation peculiar to money: one axis that relates to things, and one that relates to time. Both axes are of course prone to speculation, directed at differential in value of stuff, as a matter of exchange, and directed at differential in value over time, as a matter of contractual obligation, respectively. Both of these two dimensions of value assets have been heavily developed in the market for so-called “financial products” that has boomed since the massive liberalisation of financial activities through the nineteen-eighties and the new medial affordances provided by information technologies.

Speculation, in financial terms, is distinct from investment. Investment is a bet on future returns based on the assessment of the venture in question and the surplus value it might provide, and thus the relation to time and future value is based on a relation to things and their potential. One of the things that has happened in the financialization of contemporary economy is that the relation between the investment and the object of said investment is increasingly becoming opaque. In the “old” economy, the investor knew about the production or the property he was buying into and based his expectations for future returns on this: assessing risk and the balance between risk and interest rate. In the new economy, this direct relation to the object of investment is no longer to be discerned, instead we have an endless chain of relational patterns, through the relays and interfaces of different “financial products”. Somewhere out there an object still prevails, perhaps a fraction of a mortgage in a derelict house, but it is out of sight. In securitized credit packages, ramifying engagement to minimize risk, the actual instances of risk are blocked from sight. And decisions are made, then, not on the basis of actual risk assessment, but on the expectations about the expectations of others, etc., on the “mood” of a market rather than on insight into the matter that is traded.

The most flabbergasting symptom of this development is probably the invention of the “collateralized debt obligations”, or CDOs, one of the most efficient drivers behind the financial boom at the turn of the twenty-first century, and indeed of the notorious meltdown in 2008. To collateralize is a matter of diversification of engagement, where you don’t invest in one single asset, but instead put together a large package of assets and then slice up the package in order to invest in very small fractions of a very big number of assets, leaning again on magic qualities of big numbers, but also virtually eradicating the actual economical referents behind the assets, thus opening up for toxic mortgages or otherwise dubious ventures. These blindfolded obligations can then be submitted to betting on the financial market. As their relation to things, to actual economical assets is however unclear at best, their future value cannot really be assessed, which is why the CDOs are now being accompanied by another shrewd instrument, the CDS or the “credit default swap”, which is basically an insurance contract between two partners to yield a specified compensation if the value of the asset in question rises or drops beyond a certain limit. With this instrument, you can “hedge” your investment, as it is possible to bet on both sides of a given asset, making money if the value goes up, and making money if the value goes down. This, however, does not make it easier to actually assess the value of a CDO; but then econometrics, in guise of the 1997 Nobel Memorial Prize winning “Black-Scholes formula”, devised a mathematical model for the prizing over time of such obligations. And this model was of course rapidly integrated in the trading algorithms, making the representation of putative market reactions into the actual operator of market decisions.

Two lessons about what has happened to speculation in the era of Big Finance can be derived from this abbreviated tale: the first one is about things, how market axiomatics is deterritorializing itself and cuts loose its relation to economic assets, and the second is about time, how the market is gradually reterritorializing itself on the future as a source of gain in the present. And in both dimensions, contemporary financial capitalism can be seen as a successful attempt to re-mediate money from being a representational medium into a still more autopoetic medium.

What characterises the market of futures, CDOs and CDSs, is that the trading (and the tremendous gains it produces) functions almost without touching ground, without engaging with any specific reference, without any burdensome grabbling with economic reality. In the wake of the 2008 meltdown, this increasingly autonomous financial market was referred to as “bubble” that had eventually inflated. The bubble is an interesting metaphor because of the specific quality that it can suddenly blast, but it also highlights that what blasts is a defining membrane separating an inside from an outside. And contemporary financial speculation is precisely operating on the premises of such a separation and an autonomization of speculative agency. Autonomy means a loss of – or a freedom from – the two traditional poles of investment: the object of investment and the outcome of investment, that is, where the traditional banker exercised his judgment, weighing opportunity and risk. In his book on the 2008-crisis with the telling title Freefall (2010), Joseph Stiglitz thus diagnosed a specific blindness of the new bankers: a blindness towards agency, i.e. the actual economic activity that investments rest on, and a blindness towards externalities, i.e. the multifarious side-effects of financial dispositions. The deterritorialization of financial speculation has operated a distribution between inside the sphere of the bubble and the outside and thus replaced economic agency and externalities by parameters pertaining to the inner logic of financial circulation, those of futures, ratings, pools, algos, short-term interest rates obtained through rearrangement of placements, and of course the market value of financial labour.

Speculation eventually here is not speculative investment in something else – the production of this or that – but an operation in a closed market place with its own rules. Like in the case of big data mining, we are here looking at a transgression of speculation as it spills over into the framework that should originally have served to test its evidence. This said, however, the new autonomy of financial operations as they are being deterritorialized from their sub-stratum is strictly a one-way street, as the territory of actual economic activity remains heavily exposed to what happens inside the bubble of financialization. Speculation economy started out as a small supplement to the economy of production, perhaps even affording it by assuring rapid flows of liquidity, but throughout the latter part of the twentieth century, the bubble has grown, first steadily, then exponentially, until eventually the hierarchy of their relation was turned upside down. Henceforth, the economic activity has relied completely on the whims and stirrings inside the financial bubble. Which takes us back to the 2008 burst as a veritable clash of realities. A clash between the inner logic of the bubble, delinked from the realities of economic agency and regulated by complex algorithms built into opaque financial products, incentives of short-term interest rates and contract based bonuses. And another side a mundane reality, where very real people tend to lose their homes and jobs when one algorithm meets another.

This, however, is only half of the history about what happens to speculation in the contemporary hypertrophy of financialization. The perilous deterritorialization from things, from the matter of economic activity, is accompanied by a reterritorialization on time, on futurity. If deterritorialization is primarily operating through “securitization”, like with the CDOs mentioned above, the parallel re-territorialization takes place through “derivation”, the issuing of derivatives like the CDSs. The derivative is a bet on the future increase or decrease in value of a position, that is, in Arjun Appadurai’s succinct wording, “a bet on the risk of a defaulting debt” (Appadurai 2015, 136). With derivatives, we have taken the definitive step beyond investment as money is no longer placed in the development of infrastructure for production but solely placed on the default (or not) of any given contract, which is why, as argued by Brian Holmes (2014), we really shouldn’t think about this as a market place but rather as a casino. Moreover, the individual CDSs, the betting contracts on the default on any conceivable debt, can themselves be subjected to further betting, thus instigating an entire series of bets on bets on bets on future defaults. Which in turn really makes money flow, or as it is put, creates liquidity in the market: you can make money in the present by buying and selling contracts on possible future defaults.

Betting here is based, at Joseph Vogl has pointed out, not on an epistemology of value but on an axiology of trading, because betting is no longer based on the value of an asset but on contractually fixed expectations pertaining to future price differentials. In a certain sense, this betting is not only projecting the present onto the future, like in the case of the investors deliberation on potential future value, it furthermore projects this future back onto the present. And then strange casualties occur. What happened in 2008 was the projection of a failed future onto a present moment. Or with Joseph Vogl again: a moment “where the ghost of capital came back from its own future” (Vogl 2010, 172).

Deterritorialization and reterritorialization, mediated by securitised bonds and derivative bonds, respectively, work tightly together. The first is an instrument for amassing debt, while the second is an instrument for making money out of the possible contingencies of that debt in the future. And the more liquidity you can mobilise in the present, the more potential for creating future debt positions, while the more future debt positions you have circulating in the market, the more credit to bet on, out of which can be made additional liquidity. This double pull of modern financial speculation thus reaches beyond making hypotheses about the future; this is what the old-style investor did. Rather, it inflates future debt through securitization, and it makes more money on future debt in the present through derivatives grafted on that very debt. Unlike traditional speculation, but very much like the one thriving in the businesses of data mining, it really cannot be proven wrong (or rather, when it is proven wrong, somebody else will have to take the tab, as we have also seen in the wake of 2008). We don’t speculate on the real while we look at it through some medial machine: by now, we’re inside the machine.^[4]

According to Elena Esposito, the calculus of probability came into being in the early eighteenth century as a reaction to the increasing complexity of modernity. As societal processes became harder to assess, probability offered a second-order system that could reduce complexity by isolating salient variables and making them subject to calculation. This instrument has been endlessly refined through the centuries. But a qualitative shift might be occurring with the advent of big data, due to the sheer bigness of the operation. By now, as argued in the previous section, calculations have shifted from being an analytical apparatus to become a practical apparatus, acting directly on and interacting with the population: no longer a model of the real, but a world-making tool in its own right.

Paper money comes into being at about the same time, successfully in England, more troubled in France, with the John Law system of 1716, which immediately resulted in a major speculation crisis. Generally, two effects are assumed to follow from issu-ing money: it facilitates circulation, or liquidity, and thereby accelerates productivity, and it invests the obligation of the state in surplus growth; both these functions are clearly spelled out in Mephistopheles’ poetic parlance in Goethe’s Faust II. So from the outset paper money is speculative – resting on the assumption that the future will be able to honor the present’s betting on the state’s ability to cover its debt and on the beneficial effect of efficient mechanisms of liquidity. Hence, modern monetary policy has always had to manoeuvre these implicit twin stakes, the risk of inflation on the one hand, and the risk of stagnation on the other. This juggling has another well-documented 300 years history. What is new here is, again, pretty much about quantity. About increasing debt (extracting still more capital from our need for housing, education, health), and about the increasing efficiency of financial markets, due to mediatization through information technologies, heavily spurred through neoliberal deregulation, so that by now, economic activity is no longer serviced by the financial techniques of providing liquidity, but has itself become a source to service financialization.

3 Literature

What I have been attempting in the two previous sections is to portray a historical change in the role and function of speculation, from being a tool to administer modernity by facilitating the government of populations or the circulation of capital, to becoming an actual social logic that modulates our everyday lives and sets the premises for economic activity. A change, in other words, from being a mode of hypothetical representation to being a mode of intervention, where second-order systems, to use once again the illuminating Luhmannian term, blend into first-order reality.

My reason for venturing into these areas, indeed off the path of literary studies, and my curiosity as to the fate of speculation today, does however stem from an observation within the realm of literature proper. It has struck me that “fictionality”, a speculative mode par excellence, has become less well delineated as a feature of literature over the latest decennia. One thing is that many writers and other professionals in the literary field now challenge or even repudiate the “contract of fiction” that has characterised the epistemological and legal status of literature for centuries. But moreover, it also seems that some of the characteristics of fiction making are now spilling out into society at large, when “telling a good story” tends to become no less important than portraying reality accurately, whether in politics, branding, commercials or journalism. Perhaps the distinctiveness of fiction as a literary property is gradually being attenuated, and with it the demarcation line that has for centuries separated fiction from nonfiction, so that eventually the craft of fiction, again, becomes less a way of looking at reality but rather another way of intervening in the real.

In genealogical terms, the advent of our modern understanding of literary fiction is contemporaneous with the speculative practices of probability calculation and paper money. In his collection of essays, Farther Away, Jonathan Franzen discusses fiction as another tool for the administration of modernity, which came into being in the early eighteenth century as a specific modality of speculation:

When business came to depend on investment, you had to weigh various different future outcomes, when marriages ceased to be arranged, you had to speculate on the merits of potential mates. And the novel, as it was developed in the eighteenth century, provided its readers with a field of play that was at once speculative and risk free. While advertising its fictionality, it gave you protagonists that were typical enough to be experienced as possible versions of yourself and yet specific enough to remain, simultaneously, not-you.

(Franzen 2011, 32)

Fiction, then, as the notion becomes codified with the modern novel, is an experimental form, a medium for conjecture, for trying out, for testing hypotheses, for speculating. It is based on a contract, as Catherine Gallagher has pointed out, that combines non-reference (this is not real) with verisimilitude (this could be real):

In England, between the time when Defoe insisted that Robinson Crusoe was a real individual (1720) and the time when Henry Fielding urged just as strenuously that his characters were not representations of actual specific people (1742), a discourse of fictionality appeared in and around the novel, specifying new rules for its identification and new modes of nonreference.

(Gallagher 2006, 344)

What comes up here, in other words, is a new framing of a set of texts, a framing that invites us to deal with them in peculiar ways, as “fictions” rather than reports and renditions. Fiction, in this sense, is not a textual phenomenon, but an institutional phenomenon. It is a framework within which expressions are produced, circulated and consumed in accordance with specific guidelines. Or again, as Franzen points out, the institution of fiction is a locus where you could speculate on reality, thus a third medium to be added to the list of early modern second-order representational systems. Hence, we should probably not talk about “fiction” as an ontological category, but rather about different historical regimes of fiction. When we make our habitual distinction between fiction and nonfiction, we are doing it according to one particular regime of fiction, and when we, as we might be now, start feeling that the ancient divide seems all of a sudden no longer to apply, we might be engaged in a transformation from one regime of fiction to another.

In modern literature, there is a long tradition for playing with fictionality, which has indeed been one of the preferred ways for literary works to reflect on their own status as texts that are framed in a peculiar way, distributed in specific circuits and read with a very precise set of pre-apprehensions. What seems to have happened over the latest decennia, however, is that this kind of playful challenging of the very framework of literature has not only been multiplied, but even reaches a level where transgression is hardly any longer a point in itself, but rather an outreach for a new key to understand the relation of life and literature.

A crucial moment in the history of literature where the now habitual combination of verisimilitude and non-reference somehow did no longer apply, was the advent of witness literature. Elie Wiesel put it in this way: “If the Greeks invented tragedy, the Romans the epistle, and the Renaissance the sonnet, our generation invented a new literature, that of testimony” (Wiesel 1977, 9). The testimonial quality of witness literature almost turns the contract of fiction upside down; for one thing, it insists on referentiality, on stating something that actually happened, and furthermore, it confronts the idea of verisimilitude by bringing the unheard-of, that which would otherwise not have been testified to, into language, into the common. At the hearth of the testimony is the absolute singular, the unfathomable destiny that doesn’t have a proper designation in common language. And the task of this literature is then to invent a language for such experiences, such parcels of reality, eventually to rework and deform common language in order to bring this singularity to a possible expression. Surely, witness literature is literature, but it is not fiction according to our traditional understanding; “fictio” here refers less to fingere, to the operation of the as-if, and more to facere, the operation of producing something, developing an expression.

In the same vein, we have seen a constant upsurge of documentary methods and forms and an on-going consecration of such forays into the real, from the award of the Golden Palm to Michael Moore in 2004 to Svetlana Alexievich’s Nobel Prize in 2015. The speculative practice of creating a literary universe here refrains from taking a step back to create a model of the real under the aegis of the famous “as if” – now it rather intervenes in an on-going process of inventing and developing a way to talk about, to understand, to imagine the real in which we live.

On June 21, 2013, at the publication of the American translation of the second volume of Karl Ove Knausgaard’s My Struggle, the reviewer for The New York Times wrote: “immediately striking is the way in which fiction is born of fact,” and then thoughtfully added, “and the question whether this is fiction at all”. This question has occupied literary scholarship and the literary public to a quite remarkable degree over the last years, not only in the case of Knausgaard, but in the panoply of instances where contemporary writers in different ways transgress the ancient contract of fiction, from W.G. Sebald to Michel Houellebecq, from Marie Darrieussecq to Rainald Goetz. All the prominent theories of fiction in store have been invoked, and new sub-generic classifications have been laboriously devised, but somewhat, it seems, in vain. In the different explorations undertaken by writers like these, we are no longer dealing with new, subtle negotiations of the contract of fiction, but rather with a practice to which the contract and the divide it implies is simply becoming increasingly irrelevant. With characteristic perspicacity, J.G. Ballard already in 1995 stated:

I feel that the balance between fiction and reality has changed significantly in the past decades. Increasingly their roles are reversed. We live in a world ruled by fictions of every kind – mass-merchandizing, advertising, politics conducted as a branch of advertising, the pre-empting of any original response to experience by the television screen. We live inside an enormous novel. It is now less and less necessary for the writer to invent the fictional content of his novel. The fiction is already there. The writer’s task is to invent the reality.

(Ballard 1995: i)

“Inventing reality” could be an apt formula for the aspect of literary speculation that comes to the fore once the traditional framing of literary invention in terms of non-reference and verisimilitude is coming undone. Rather than speculating on the world, it too embarks on speculating in the world, giving voice and consistency to what is already there, visible but unseen, and thus taking part in the bigger discursive operations of world-making, to use Nelson Goodman’s suggestive term. This is not, to be sure, a brand new feature of literature, rather one of the most powerful properties in its long history, but it has to some degree been held in check by the idea of fictionality and the ontological divide between what is real and what is invented. But in a situation where discursive world-making strategies as the ones mentioned by Ballard are on the rise and where all kinds of speculation are weaved still more tightly into the fabric of social life, the idea of leaving the notion of fictionality behind altogether is becoming increasingly tempting.

In February 2003, secretary of defence Colin Powell addressed the UN Security Council to present his findings about the alleged Iraqi weapons of mass destruction and declared: “Every statement I make today is backed up by sources, solid sources. These are not assertions. What we are giving you are facts and conclusions based on solid intelligence”.^[5] This was not the case. He was probably not outright lying, though, but he demonstrated the trickiness of facts and the importance of a solid discursive scaffolding to uphold whatever findings one might be able to produce. This predicament was taken up by Jacques Rancière in a talk at the opening of the 2011 Venezia Biennale on “The State of Things”:

Strictly speaking, the state of things is a fiction. A fiction is not an imaginary tale. A fiction is the construction of a set of relations between sense and sense, between things that are said to be perceptible and the sense that can be made of those things. […] In that sense, a “state of things” is a form of what I have proposed to call a distribution of the sensible: a set of relations between the perceptible, the thinkable, and the doable that defines a common world, defining thereby the way in which, and the extent to which, this or that class of human beings takes part in that common world.

(Rancière 2013)

Here, the notions of what is real and what is fiction and the defining opposition between them are irremediably being undone and replaced by the outline of a new strategic field of discursive agency. If this is indeed the field contemporary literary practices find themselves situated in, it is understandable that the sacrosanct modern understanding of fiction as an experimental second-order system, which has otherwise served us well for some 300 years, seems to have become untenable. Deprived of its defining framework, the production of fiction is no longer about testing speculative models, as described by Franzen. Literature becomes still less about fiction and still more, then, about fabulation, producing “a set of relations between the perceptible, the thinkable and the doable that defines a common world”. Such writing practices are no longer hedged by the caveat of “being just fiction”. They stand up to their role of intervening in the textual landscape of the contemporary social imaginary, in our shared and individual modes of self-fashioning, and take their place as something we might call – literature beyond fiction.