In this essay, I view the eighteenth-century novel as part of a cultural system that worked to validate Enlightenment canons of knowledge by dynamically linking the realms of science and fiction in the very process of setting them in opposition. In contemplating the historical particulars of this always mobile counterpoise, I focus on a realignment that occurred around 1750, when the guarantee of factuality in science increasingly required the presence of its opposite, a manifest yet verisimilar fictionality in the novel. My method is purposely to erase boundaries among specialized senses of words that, even today, remain central to our protocols of knowledge in order to suggest how freely the problematics they engage jostled with one another during a period when the modern disciplines and professions were emerging in correlation with new divisions of knowledge. I have ignored the tiny parentheses that followed headings--e.g. HYPOTHESE (Métaphysiq.)--in the dictionaries and encyclopedias that played a crucial role in marking out these divisions and in linking words to disciplines. Parenthetical allocations in Denis Diderot's and Jean le Rond d'Alembert's Encyclopedia--for instance, dividing VERISIMILITUDE into three specialized articles marked "metaphysics," "poetry," and "painting," or designating PROBABILITY to "philosophy," "logic," and "mathematics" while treating related issues in the mimetic arts under a selection by a poet about FICTION--block the crosstalk that I study here. These tiny parenthetical marks in the semantic field are traces of fierce struggles in which different meanings were assigned to fields as part of the process that defined and maintained their disciplinary autonomy. Watching words in this way lets me propose an affinity in difference between Enlightenment science and the novel.[1] I also regard the eighteenth-century trends described here as elements in a long term set of cultural transactions concerning the maintenance and ownership of impartial discourse because public exchange aiming toward consensus was the medium that enabled the modern sense of objectivity and constructed and stabilized its terminologies. These include not only the concepts of "hypothesis" and "fiction" that this essay explores but related ones like "verisimilitude," "probability," "fact," "history," and even "truth" that remain vital in present day philosophy of science.[2] For, only through discussion and contestation in public did the array of terms treated here come to signify differently than they had before the Scientific Revolution and to assume different interrelationships. The article in the Encyclopedia on "hypothesis" (1765) that figures below, for instance, at once impartially reflected a positive consensus about the value of hypotheses in science emerging at mid-century and defined an arena of debate by cordoning off both Cartesians and Newtonians as extremists in a manner reminiscent of Denis Diderot's attack on both schools as abstractionist and anti-empirical in his pornographic novel, Indiscreet Jewels (1748). In doing so, the article also implicitly engaged the sharp division that opened up around mid-century between two forms of science. The first was a speculative science theorized by Diderot, for whom hypotheses were a guiding value and function when anchored empirically in experiments that provided continual feedback from observation. The second was an aggressive hyper-inductivism typified in England by Thomas Reid, which also was experimentalist but neo-Newtonian in its absolute rejection of hypotheses and its ridicule of them as dangerous fictions.[3] The Encyclopedia was a symptom of the Enlightenment's embrace of new ways of formulating knowledge based on empirically oriented critical communication, which opposed itself to the abstract system- making characteristic of late medieval Scholasticism and early modern metaphysics.[4] Consensus, the essential yet mobile reference point that defined critical discussion aimed at the progressive improvement of knowledge, is the hallmark of modern enquiry as epitomized in Kant's essay, "What is Enlightenment" (1784). Here, of course, I also have in mind Jürgen Habermas's description of the eighteenth-century public sphere as a scene of communication freed from the constraints of courtly hierarchy and a priori thinking: within this arena could thrive a public science in which lectures and demonstrations in coffee houses and other sites took place on a continuum with formal proceedings and in which could develop descriptive techniques for making scientific experiments vividly present in published form.[5] Empiricism, the experimental method, and an emergent sense of "science" as progressing through accumulation and refinement, all fostered the ideal of impartial knowledge as a product of rationally moderated interactions among such problematics as: (1) theories of observation; (2) the derivation of "facts" from nature directly via the senses and indirectly through their extension with instruments; and (3) the practice of inductive thought, which arrived at synthesis by methodically breaking problems down into sequential, empirically definable parts and judging validity or "truth" according to coherence and completeness of sequence and probabilistic weighing of evidence. Although derived from mathematics and statistics, such methods extended as far as rules of evidence in law and even to proofs of God's nature in theology[6] The "new" novel of the earlier eighteenth century may be described an institution of the Enlightenment not only because it partook, to greater or lesser degrees, in what Webster's dictionary calls "a philosophic movement . . .marked by questioning of traditional doctrines and values, a tendency toward individualism, and an emphasis on the idea of universal human progress, the empirical method in science, and the free use of reason" but, more significantly, because the novel operated within the Enlightenment model of critical communication. The novels of Daniel Defoe, Samuel Richardson, and Henry Fielding, for example, not only pretend to offer densely particular, virtually evidentiary accounts of the physical and mental circumstances that actuate their characters and motivate the causal sequences of their plots, but these novels also attempt to frame the subjectivity of their characters within editorial objectivity, as in Defoe and Richardson or narratorial objectivity, as in Fielding. These novels share a way of representing the world and kind of verisimilitude that Shapin and Schaffer in Leviathan and the Air Pump call "virtual witnessing," by which they mean the rhetorical and visual apparatus for communicating scientific experiments to a public and convincing that public of their authenticity. Shapin and Schaffer cite, for instance, the hyper realism of diagrams in which additions of fine detail and unneeded illusionism work to satisfy and convince an audience.[7] Samuel Johnson defined "verisimilitude" simply as "probability."[8] These synonyms are not directly the subject of this essay but they are constant referents of the concepts that are, and they recur continually in discussions both of scientific method and novelistic practice. These axes of terms are not accidental. From a philosophical perspective, all hypotheses are fictional but the most serviceable (those that appear to yield scientific truth) are the most probable given the evidence and the prevailing rules of verification--which is to say the most verisimilar. Paradoxically, the best hypotheses are those that seem the most like truth. Yet, their very plausibility also renders them suspect because they are products of imagination that are logically indistinguishable from poetic fictions. David Hume held that "vivacity" in "poetry . . . never has the same feeling with that which arises in the mind, when we reason, tho' even upon the lowest species of probability" because ideas in manifest fictions rely on imagination rather than on distinct impressions.[9] He cannot say the same of hypotheses, which early fall victim to his deconstruction of most causal connections as fictional. Hume struggles valiantly to feel the difference between fiction and reality but, on his own skeptical account, verisimilitude is essential to our apprehension of both because both rest upon probable inference and upon the fiction of causal continuity.[10] As Hume says, "all knowledge resolves itself into probability, and becomes at last of the same nature with that evidence, which we employ in common life. . . . In every judgment, which we can form concerning probability, as well as concerning knowledge, we ought always to correct the first judgment, deriv'd from the nature of the object, by another judgment, deriv'd from the nature of understanding. . . . As demonstration is subject to the control of probability, so is probability liable to a new correction by a reflex act of the mind, wherein the nature of our understanding, and our reasoning from the first probability become our objects." (Treatise, 181-2) Unlikely as the pairing may at first appear, Hume's simultaneous skeptical and commonsensical--now you see it now you don't--position about reality and fiction shows an affinity with Samuel Richardson's about the kind of conviction he wished readers to experience in the letters that make up Clarissa (1747-48): "I could wish that the Air of Genuineness had been kept up, tho' I want not the letters to be thought genuine; only so far kept up, I mean, as that they would not prefatically be owned not to be genuine: and this for fear of weakening their Influence where any of them are aimed to be exemplary; as well as to avoid hurting that kind of Historical Faith which Fiction itself is generally read with, tho' we know it to be Fiction."[11] Fictions, be they hypotheses or novels, yield a provisional reality, an "as if," that possesses an explanatory power lacking in ordinary experience. Science needs such separation of its findings and procedures from the ordinary, and explanatory hypotheses have compelling power as stories that approximate truth, but the scientific disciplines cannot tolerate the imputation of fictionality. This is why induction must emerge as the opposite of hypothesis in scientific method--because it attains or seems to attain the independence from the fictional.[12] Even the novel shares the impulse toward formal proof so strongly voiced in the embrace of induction by scientific theorists inspired by Newton. Fielding declares, for instance, "in reality, there are many little circumstances too often omitted by injudicious historians, from which events of the utmost importance arise. The world may indeed to considered as a vast machine, in which the great wheels are originally set in motion by those which are very minute, and almost imperceptible to any but the strongest eyes."[13] But while novels may present attributes of inductive proof, they cannot attain such status and find themselves recurring to other models like the juridical. Thus, Richardson, directly linked circumstantiality, probability, and legal proof when he wrote (after the 1749 publication of Fielding's Tom Jones), that "the Probability of all Stories told, or of Narrations given, depends upon small Circumstances; as may be observed, that in all Tryals for Life and Property, the Merits of the Cause are more determinable by such [details], than by the greater Facts; which usually are so laid, and taken care of, as to seem to authenticate themselves."[14] In point of both thematic exposition and narrative strategy, these novels force readers into the position of neutral observers arriving, probabilistically, at judgments based upon the weight of available facts and reasonable inferences. Although these works continue to echo plot forms and incidents traditional to the fantastical seventeenth-century romances their prefaces ritual condemn, the concern expressed by their authors to distinguish their enterprise from that of earlier prose fiction signals the reformulation of generic traits and of cultural function that transformed the novel during the eighteenth century. Richardson was tacitly acknowledging that Tom Jones decisively marked the success of Fielding's effort to find formal means to legitimate novels as objects of and forums for critical discussion by bringing to the foreground issues of evidence and probability. Yet, given the moment of its publication in 1749, Tom Jones also must be viewed as a monument to a fading ideal of apparent historical factuality as the basis for impartiality in the genre of the "novel" because Fielding's work also embodied a turn, characteristic of mid-century, toward a manifest fictionality that tends to undercut the novel's own claim to simple factuality as a basis for its truth. Thus, Michael McKeon can argue in The Origins of the English Novel, that the novel at mid-century was giving up claims to embodying literal, historical truth in favor of claims to manifest higher truths through the transparently fictional construction of specialized versions of the "real" world within strict canons of physical, temporal, and psychological probability.[15]
I. Questions about fictionality are crucial to the period's theories of both literature and science, yet a certain denial of fictionality marks both the earlier eighteenth- century novel and early science. Two famous, nearly concurrent instances from Isaac Newton and Daniel Defoe, can serve as launching points for this discussion. Newton, in the "General Scholium" added to the second edition of his Principia in1713, wrote the much disputed words "hypotheses non fingo," which early were translated as "I do not frame hypotheses" and, more recently, have been accepted as meaning "I do not feign hypotheses." The context includes an attack on Cartesian celestial mechanics and an assertion of Newton's sincere if unprovable belief in God as all-present cause:
The opposite of "hypothesis" for Newton was inductive proof, which found its basis in his natural philosophy on observational and experimental fact and yielded valid generalities about nature. Newton's disputed phrase is oriented both to the issue of hypothesis versus fact or truth, thus defined, and to the issue of feigning--for the Latin root of the word "fiction" is "fingo" and, as we shall see, it is no accident, given sound ancient precedent, that Newton linked this concept with that of hypothesis. Defoe, who like Newton believed his representations embodied the higher truth of God's being, took on the role of Editor in the Preface to Robinson Crusoe of 1719, in order to assert the standing of his text as "a just history of fact." He instantly defined the opposite of "fact" by declaring that his story is without "any appearance of fiction in it." That same year, he maintained this stance in the Preface to the Farther Adventures of Robinson Crusoe, with the slight concession that the "just application of every incident, the religious and useful inferences drawn from every part. . . must legitimate all the part that may be called invention or parable." The critic Charles Gildon shortly attacked both works for their implausibility not only as "fact" but even as "fable," cataloging one implausibility after another and in the process revealing criteria for verisimilitude so stringent that he would have required the character Xury to speak Arabic rather than English.[17] Though from different realms, both Gildon's literalist response to Robinson Crusoe and Newton's ever more intransigent statements against hypotheses added to the Principia during the 1720s, may be viewed as indices of a crisis concerning the nature of factuality and verisimilitude that Defoe tried to bridge using terms like "parable" and "allegory." Both Newton and Defoe in their different ways are forced to back away from the absolute stance. Newton, in a 1726 addition to his "Rules of Reasoning in Philosophy" in the Principia, said:
Truth, it turns out, is a matter of approximation. Speaking as Editor, Defoe retreats also--even within his very own sentence, asserting that Robinson Crusoe's powers of "improvement. . . will be the same," whether fact or fiction, since in either case it works "as well as to the diversion, as to the instruction of the reader." Analogously, Defoe questions the factuality of "fact" in Robinson Crusoe's Serious Reflections (1720), his defensive sequel justifying the original novel as true in the larger sense of parable or allegory, declaring that "nothing is more common, than to have two men tell the same story quite differing one from another, yet both of them eye- witnesses to the fact related." In a strange kind of consonance with Newton's "hypotheses non fingo," Defoe declares that "this supplying of a story by invention. . . is a sort of lying that makes a great hole in the heart."[19] Newton's propositions, it seems, must withstand hypotheses while partaking in the partiality of being "very nearly true"- - and Defoe's facts slide easily, through the process of narration, into untruths that erode self- certainty. In both cases, narration symptomizes the disease of fictionality. Narration bears with it the infection of fictionality. Newton sets the feigning of hypotheses, which spring from the imagination, opposite the accuracy of propositions yet these, too, are only relatively accurate, very nearly true, and liable to exceptions. Their epistemological standing is different from hypotheses yet they share the extension in time fundamental to narrative and are threatened by its tendency toward fictionality--even, as Defoe pointed out, in eye witness accounts. Hans Vaihinger argues in The Philosophy of 'As if, that "the Greeks made no clear-cut distinction between hypothetical and fictional assumptions."[20] Indeed, the very concept of hypothesis, though urged toward different status by Copernicus and others during the Scientific Revolution, traditionally referred in antiquity to the narration of specific facts and particular cases treated by rhetoricians or to a plot outline. According to Wesley Trimpi, citing the Nicomachean Ethics (3.3.9-13), "the philosophic-scientific function of 'hypothesis' in a mathematical construction begins simultaneously with the choice of an 'hypothesis' in the dramatic sense of a plot outline."[21] In Quintilian, too, the relationship between hypothesis and fictional argumentation and narration remains intimate: "when I speak of fictitious arguments I mean the proposition of something which, if true, would either solve a problem of contribute to its solution." Finally, as Trimpi shows, "the use of `hypothesis' in a literary context [was] not restricted to summary or outline, but [gradually came] to refer to fiction itself or rather to [that] historically important type of fictional narrative, the verisimilar."[22] Newton's theoretical condemnation of hypothesis took these traditional associations as its premise and they help to explain his attempts to disentangle himself. Despite Newton's late theoretical declarations, his attitude toward hypothesis turns out to have been remarkably unstable across his career. In practice, Newton himself did form hypotheses and often in his works had gone far as to call them just that. Historically, as the context of the 1713 "General Scholium" makes clear, his condemnation was aimed at the extravagant use of explanatory hypotheses without experimental reference that marked Cartesianism and Leibnitzianism as he viewed them. With this context in mind, then, the key word in Newton's phrase becomes "fingo," rather than "hypotheses," and his statement is reoriented away from the issue of fact or truth versus hypothesis to the issue of feigning: that is, to the issue of manifest, empirically unanchored, fictionality. This Newtonian tendency to suspend or erase the speculative function--including its uncomfortable associations with verisimilar narrative, as opposed to the factual truth of demonstration in both its analytic and physical senses--continues with surprising strength in both English and French science. Joseph Priestley declared, for instance, in a lecture of 1761 at the Warrington Academy that, "All true history has a capital advantage [over] every work of fiction" because "works of fiction resemble those machines which we contrive to illustrate the principles of philosophy, such as globes and orreries, the use of which extend no further than the views of human ingenuity; whereas real history resembles the experiments by the air pump, condensing engine and electrical machine, which exhibit the operations of nature, and the God of nature himself."[23] In France, Voltaire had styled le Père Louis-Bertrand Castel, a Jesuit popularizer of mathematics and science whom he loved to hate, as "the Don Quixote of mathematics," while the 1739 revision to his essay in the Philosophical Letters, "On the System of Attraction, " condemned Descartes because, in his desire "to create a universe, he made a philosophy the way one makes a good novel: everything seems verisimilar, and nothing is true."[24] Questions about fictionality are crucial to the period's theorization not only of literature and science but to its general epistemology. The Encyclopedia devotes more than three folio pages to an article on poetic "fiction" by the poet Jean-François Marmontel (1756). The more technical aspects of the topic, "fiction," are managed chiefly through cross references and in other articles but Marmontel allows himself a paragraph praising the new "esprit philosophique" as condemning "extravagant fictions, while embracing poetic fictions that respect probability and verisimilitude because the new philosophy, which "observes, penetrates, and unfolds nature. . . alone is capable of appreciating imitation because it alone knows the model." Discussions dealing with the matter of fiction appear in the main text and supplements on topics such as "fact" (by Diderot), "experiment," "induction," "unbelievability" (by Diderot), "observation," "phenomenon" (by d'Alembert), "probability," "verisimilitude" (possibly by Diderot). The Encyclopedia article on "hypothesis," by the mathematics teacher Abbé de La Chapelle, also deals centrally with the question of fiction.[25] It is a virtual panegyric about the centrality of hypotheses to scientific progress and specifically contrasts the orderly, empirically oriented usage of hypotheses in the scientific method with their fabulous employment by the followers of Descartes. He presents Newton's condemnation as an overreaction to the abuse of hypotheses and urges what we may take as the common sense view of the mid- eighteenth- century: namely, that the process of hypothesis is necessary to scientific research which, by its nature, must stretch over long periods and must be conducted by many different people "before attaining a certain perfection." Hypotheses, he says, need to have a place in science because they "give us new points of view," an idea that seems to point to something like Nelson Goodman's "world making."[26] And yet he confidently asserts that hypotheses can be falsified and recognized as "fictions" if they are "made up of empty terms or ones that lack fixed and determinate ideas, if they explain nothing, [or] if more difficulties follow from them than they resolve, and so forth." This proto-Kantian notion of the perfection of knowledge through public circulation interestingly considers hypothesis as a medium of critical communication. The Encyclopedia article epitomizes in everyday language a practical yet theoretically treacherous understanding of the relationship among fictions, hypotheses, and verifiable findings that has informed countless, more technical discussions by "realists" from Newton to Vaihinger.[27] Diderot expressed it metaphorically in his Thoughts on the Interpretation of Nature (1754):
Throughout this work, Diderot strives to define the practice of objectivity as a dialectical fusion of observation and creative speculation, which yields probable, verisimilar representations of nature. As Diderot's metaphor implies, such a fusion transcends tendency of individual subjectivity to fantasy or fiction by employing sensory feedback--an effect that is amplified by critical communication. Even Hume, whose far more skeptical account of causality as entirely mental has devastating consequences for the distinction between fictions and hypotheses, remained convinced that he could feel the difference between manifest fictions such as those of literature and logically fictional real entities because manifest fictions rely on imagination rather than on distinct impressions, which Hume says are physically closer to reality.[29] But even for Hume facts, unless of the most rudimentary physical or historical kinds, are verisimilar representations based on consensus. Madame du Châtelet, an ardent defender of hypothesis making in science, reveals with perfect clarity groundless grounding that was usually hidden in defenses of hypothesizing as the path to truth--that is, the subjectivity underlying the "objectivity" of science:
At the end of Madame du Châtelet's statement, the divergent ideas of scientific and moral certainty suddenly converge paradoxically and peep slyly through the standard disguise of probabilistic rhetoric. What I want to consider here in thinking about the novel together with the debate over the fictionality of hypotheses is not the merit of eighteenth-century arguments about the scientific validity of hypothesis but rather the significance of the debate's persistence over such a long period from the seventeenth century onward. This period coincides with the development of the modern novel, and eighteenth-century proponents of the genre and its critics alike, obsessed with questions about its standing as a form of knowledge, recurred again and again to concepts that haunt both the scientific discussion and debates like Defoe's with Gildon: "fact," "fiction," "probability," "verisimilitude," "truth" are the common terms.[31] I suggest that the novel, in tandem with the Enlightenment debate under review here, partook of a crisis precipitated when the Scientific Revolution disrupted the old continuity between the hypothetical and the fictional, which Vaihinger and Trimpi document, yet could not guarantee or ultimately underwrite the difference because both are based in "cases"--that is, in causal and narrative sequences--and both share basic technologies of world making and sense making: they are not just features of logic and science or of fantasy and literature but of both. Yet in the emergent discipline of science, hypotheses, which were central to the experimental method, had to be vaccinated against fictionality. The novel, while in the main sharing verisimilar reference with empiricist science, responded to the crisis by abandoning claims to literal, historical fact of the kind Defoe had worked so strenuously to maintain and, by asserting its own manifest fictionality, strove, as Michael McKeon and Catherine Gallagher suggest, toward the representation of higher truths and toward a more intense emotional identification between readers and novelistic fictions. This novelistic occupation of the terrain of fiction then could ground the factuality of experimental science, oddly freeing it to encompass double tendencies of which the literalist inductionism of a Reid and the more imaginative versions of fact in Buffon are extreme instances.
II. I want to propose, then, that the new novel of the eighteenth century can be regarded as a guarantee of difference. It shares attributes with the hypothetical such as narrativity and verisimilitude yet its fictionality is manifest- - even paradoxically underlined through the insistence upon factuality that we have noticed the early novel. By mid- century, however, as Catherine Gallagher shows--especially with regard to Charlotte Lennox's novel The Female Quixote (1752)--the novel's attraction was increasingly based upon the readerly identification that manifest fictionality strangely enabled.[32] As Anna Laetitia Barbauld and her brother John Aikin wrote in 1773,
What happened was a specialization (running in tandem with the period's great division of knowledge) whereby manifest fictionality underwrote the factuality of science. Just before mid-century, Fielding had been poised strangely between old world of the novel as fact and the new world of the novel as fiction: his essay about verisimilitude, which makes up the introductory chapter to Book VIII of Tom Jones, attempts theoretically to protect the new novel against charges of unbelievability. These charges would arise from what Defoe called "falling into fiction." Fielding insists that the novel must not deal in the "possible" but must limit itself strictly to the "probable." As his chief modern editor, Martin Battestin, notes, he was the first consistently to have defined the genre of the novel, with reference to Aristotle, as centrally concerned with probability.[34] Yet in practice he was already of the avant-garde in relation to his wholesale abandonment, in the actual telling of the story by an intrusive narrator, of the apparatus of apparent factuality that surrounds Defoe's and Richardson's novels. By 1773, the trend toward viewing novels as fictional appeals to subjectivity had become fully articulate. By 1759, at the end of the first installment of Tristram Shandy, Laurence Sterne could declare that "it is in the nature of an hypothesis, when once a man has conceived it, that it assimilates every thing to itself." He goes on to challenge the reader to "conjecture upon it, if you please. . . . raise a system to account for the loss of my nose by marriage articles,---and shew the world how it could happen, that I should have the misfortune to be called TRISTRAM, in opposition to my father's hypothesis, and the wish of the whole family. . . . but I tell you before-hand it will be in vain."[35] Sterne is purposely dealing in the improbable everyday world where fact is stranger than hypothesis or fiction. Sterne's is the world of the "possible," a terrain that Fielding tried to forbid novelists. Following classical literary theory, Fielding allotted the "possible" to historians relating proven, thoroughly witnessed fact (which can compel belief through incontrovertible evidence), while ceding the "marvelous" to poets as tellers of miraculous events that are credible only to audiences already convinced of their truth through belief in myth or religion. For Fielding, the "probable" had been the only domain in which a novelist could move with assurance because he sought to engage readers in an economy of rational exchange, which he summed up with the word "judgment." Five years after Tristram's debut, Horace Walpole, the inventor of the Gothic novel, could formalize the turn toward manifest fictionality in the first Preface to The Castle of Otranto (1764): "If this air of the miraculous is excused, the reader will find nothing else unworthy of his perusal. Allow the possibility of the facts, and all the actors comport themselves as [real] persons would do in their situation." Interestingly, in the anonymous first edition where he poses as editor and translator of an old manuscript, Walpole also echoes previous novelistic convention when he declares, "Though the machinery is invention, and the names of the actors imaginary, I cannot but believe that the groundwork of the story is founded on truth. The scene is undoubtedly laid in some real castle." The wildly paradoxical program that Walpole outlines in his Preface to the second edition, where his initials give away his authorship, says much about the mid-century emergence of the novel as the simultaneous site of manifest fictionality and ostensible probability. Otranto, he says,
Walpole aspires to a kind of novel unmistakably fictional in its premises yet tethered to experienced reality by psychological realism and by the canons of probability. The common ground with science is, of course, that of the probable. The probable is the ground where scientific hypothesis meets fictionality; the probable is the point around which science and the novel rotate in complementary orbit, the meeting point at which fact can, apparently, be separated from fiction. Yet, Walpole's words also point to the duplicity of the probable--to its instability as a ground. For "the rules of probability" in fact leave "the powers of fancy at liberty to expatiate through the boundless realms of invention"--which is to say, fiction. We return, then to Newton's objection to feigning and to his attempt to refuse both hypothesis and fiction. For, following Friedrich Kitter, I might call the system I am describing an "Aufschreibesysteme" or "discourse network," in which all parts are systemically interconnected.[37] Scientific truth may be guaranteed by the cultural elaboration of a domain that shares certain of its protocols yet is manifestly fictional, but then science cannot fully escape implication in that very fictionality. Corresponding to the specialization in which the novel's fictionality worked to underwrite the factuality of science, whether in the Encyclopedia or in the culture at large, we find a profound relativization of narrative in science- - through the insistence upon hypotheses as necessary catalysts that structure inquiry but theoretically are not part of knowledge. This trend was accompanied by what Alexander Welsh has recently designated as "strong representations." From the mid- eighteenth century, as Welsh says, "narrative consisting of carefully managed circumstantial evidence, highly conclusive in itself and often scornful of direct testimony, flourished nearly everywhere- - not only in literature but in criminal jurisprudence, . . .natural religion, and history writing itself."[38] One thinks in this regard of Jean-Jacques Rousseau, though Welsh does not cite him. Early in the Discourse on the Origin and Foundation of Inequality (1755), Rousseau declares, "Let us begin, therefore, by laying aside all facts, for the do not affect the question. The researches in which we may engage on this occasion are not to be taken for historical truths, but merely as hypothetical and conditional reasonings better suited illustrate the nature of things than to show their true origin, like those systems that our naturalists daily make of the formation of the world."[39] According to Welsh, the powerful causal narrative structures characteristic of strong representation, "claiming to encompass more than can ever be experienced at first hand," invaded even natural science itself- - fresh evidence, I would argue, for the instability of the very edifice of scientific factuality that the novel may be seen as underwriting.
III. I conclude by reflecting on the double current of speculation in the novel and science that emerged in England and France just past mid-century. I already have noted that Lennox, Sterne, and Walpole are crucial for an understanding of this turn. An analogous figure in French science was the great mid-century naturalist George-Louis Buffon, a friend of Diderot who was attacked from every quarter for elaborating theories without supporting facts and for shaping "facts" to support his speculations. Probably, he was Rousseau's target in the passage quoted above. Buffon's work in natural history, biology, and geology, as described by modern historians of science like Virginia Dawson, does epitomize a burgeoning of speculative science at mid-century.[40] It would be easy to link Buffon to the turn to toward manifest fictionality that I have pointed to in the novel by suggesting, for instance, that the discourse network in which the novel becomes a domain of striking fictionality supporting the factuality of science and concealing its fictionality by something like relative contrast, also produces discursive space for scientific speculation. This probably is so, but contrary to Buffon's practice as viewed by contemporary critics and later historians, his own framing of arguments against hypothesis-making shows instead, or in addition, that he followed the classic strategy of attempting to maintain the scientificity of his work by associating hypotheses with fictionality and limiting their relevance to his findings, which he stringently contrasted with hypotheses and called "theories." Following Liebnitzians like Christian Wolff and Madame du Châtelet, he categorically rejects hypotheses that operate as abstractions: abstractions may be useful heuristically but science ought, instead, to be founded on the observation of concrete physical fact. Such observation yields real knowledge in the form of "théories physiques":
Although Buffon allows hypotheses, he refuses those that are disguised statements of belief or imply answers about cause and effect--cause and effect being regarded, in the spirit of Hume, as fictional constructs. As he says, introducing his vast Natural History (publication begun, 1749) in words that could as well come from Diderot or, indeed, constitute a theory of verisimilitude in the novel:
Later in the Natural History, he allows hypotheses that answer the question "how" but never those that answer the question "why."[43] Only hypotheses answering the question "how" can point to "théories physiques," the buried treasure of Buffon's science. In trying to stake out factual ground for his "théories physiques" by pushing hypothesis toward fiction, Buffon serves as the mirror reverse of Walpole's attempt to free a space for fancy by hewing to strict probability of action, characters, and psychology within initially fantastic fictional premises. Buffon is rejecting fantastic premises--which for him are concealed in hypotheses that function as abstractions or that attempt to answer the question "why"--and insisting upon strict analysis of observable physical fact in order to free a space for strictly controlled speculation in the form of hypotheses that answer the question "how." Fantasy opens space for circumstantial and psychological realism in Walpole's novel--the space within which he develops techniques for the first sustained third-person (impartial) narration of consciousness in English. "Fact" opens a space for speculative thought experiment in Buffon's Natural History.
IV. This essay has concerned ways in which eighteenth-century science worked to assume the mantle of factuality as the novel became increasingly the domain of manifest fictionality. Yet, while moving toward the positions of absolute opposition they occupy today, both joined in the larger project--seen so clearly in my quotation from the introduction to Buffon's Natural History--of defining "truth" in terms of verisimilitude or probability. Another way of putting this would be to say that they defined the most accurate representations of reality as those that contextualized empirical, sense-based "facts" by arraying them in probable explanatory networks--be these experimental cases and reports or manifest fictional narratives. Hypotheses were crucial to this mode of representation but also sites of crisis in the theory of verisimilitude because they were fictitious in essence but could become true if verified through the assimilation of enough empirical data. The contradictory nature of hypotheses produced paradoxical symptoms across the period I have described. For, against the backdrop of a shared commitment of the two realms to the system of verisimilitude, the novel began to ground the claim of science to be non-fiction by becoming increasingly the domain of manifest though probable fictionality. Then, in a doubly paradoxical twist at mid-century, as I propose, the move toward verisimilar yet "possible" fiction that occurred with Lennox, Sterne, and Walpole participated (relationally not causally) in a turn toward the more imaginative or speculative variant of "truth" advocated by Diderot and Buffon.[44] What is the long- term place of the phenomena I have been describing? The guarantee of scientific factuality through manifest yet verisimilar fictionality that I have been discussing finds a counterpart in the emergence across the nineteenth century of broader cultural forms in which probabilistic thinking dominates, such as Welsh's strong representations in legal rules of evidence and in religious thought. The rise of modern disciplines in the social sciences is well-underway by the 1830s. Terminological change during the same period suggests completion of the process of polarization that I describe above since, according to the Oxford English Dictionary, the term "novel" became synonymous in English with "fiction" in 1871. In fact, the first major instance of this usage may well have been in John Dunlop's The History of Fiction, whose his subtitle suggests that the new mapping of the terrain already had been accomplished: Being A Critical Account Of The Most Celebrated Prose Works Of Fiction, From Earliest Greek Romances To The Novels Of The Present Age.[45] It is hard to avoid the conclusion that fictionality plays a role in certifying reality even today, when long dead "real" stars like Humphrey Bogart or Marilyn Monroe can sell products of our moment on television. And when, despite infinite permutations and sophistications over the past 300 years, debates about the truth status of scientific representation are still being fought out within polarities built around terms like "realism," "verisimilitude," and even "approximate truth." Whether one notices the persistence of realism in popular and pulp fiction- - and this despite any number of twentieth- century avant- garde deconstructions- - or constant accounts in the New York Times about the rampant interfoliation of "reality" and "virtuality," fact and fiction, past and present on the contemporary scene, the questions remain similar to those raised in my contemplation of the novel and the Enlightenment. At least in the Western cultural system, are not fact and fiction, reality and verisimilitude, proposition and hypothesis, truth and narrative, inseparably bound functions of one another?
Throughout this essay, the word "science" carries its customary modern English meaning of "natural science." Although this usage is anachronistic, it is common and allows me to suggest affinities between Enlightenment and modern science. Relationships between the eighteenth-century English novel and science are not commonly discussed. Ian Watt alludes to Newton just once in The Rise of the Novel: Studies in Defoe, Richardson, and Fielding (Berkeley: University of California Press, 1957), p. 24, linking him with Locke in a discussion of the reorientation of temporality during the later seventeenth century. Changing ideas about factuality and truth in the novel are concerns of Barbara J. Shaprio's Probability and Certainty in Seventeenth-Century England (Princeton: Princeton Univeristy Press, 1983), a book that addresses shifts in the definition of knowledge during the Scientific Revolution . Lennard J. Davis's Factual Fictions: The Origins of the English Novel (New York: Columbia University Press, 1983) does not concern science but relates ideas about factuality in emergent newspaper culture and the law to the rise of the novel. More recently, Michael McKeon, in The Origins of the English Novel, 1600-1740 (Baltimore: The Johns Hopkins University Press, 1987), and J. Paul Hunter, in Before Novels: The Cultural Contexts of Eighteenth-Century English Fiction (New York: W. W. Norton, 1990), have opened questions about truth claims in the Scientific Revolution and in the early novel but the terms "hypothesis" and "fiction" in their technical, philosophic usages appear in neither book. Hunter's "Robert Boyle and the Epistemology of the Novel," Eighteenth-Century Fiction, 2 (1990), 275-291, focuses on the relationship between the novel and works by Boyle like Occasional Reflections (1665) that involve meditation on observed fact. I consider the novel in relation to Hume's and Bentham's philosophic treatments of fiction in Imagining the Penitentiary: Fiction and the Architecture of Mind in Eighteenth-Century England (Chicago, University of Chicago Press, 1987), as does Leo Damrosh in Fictions of Reality in the Age of Hume and Johnson (Madison: University of Wisconsin Press, 1989. Robert Markley treats a number of related issues in their religious context in Fallen Languages: Crises of Representation in Newtonian England, 1660-1740 (Ithaca: Cornell University Press, 1993. In Wolfgang Iser's fascinating book The Fictive and the Imaginary: Charting Literary Anthropology (Baltimore: The Johns Hopkins University Press, 1993), Newton is twice mentioned and his statement against hypothesis is quoted but only in the context that places Newton at the end of the Renaissance tradition of viewing "fiction" negatively. Iser never discusses the term "hypothesis." And so little does the novel figure in Iser's book that Richardson is not mentioned at all, while Defoe and Fielding each appear only once, in contexts that have nothing to do with the genre of the novel per se. Most recently, Everett Zimmerman's The Boundaries of Fiction: History and the Eighteenth-Century Novel (Ithaca: Cornell University Press, 1996) does not deal with science but does treat some other issues that come up here. [2] Discussion of the issues surrounding these and related terms occur throughout two recent guides that have been helpful in the preparation of this essay. The Philosophy of Science, ed. Richard Boyd, Philip Gasper, and J.D. Trout (Cambridge, MA: The MIT Press, 1997) and Companion to the History of Modern Science, ed. R. C. Olby, C. N. Cantor, J. R. R. Christie, and M. J. S. Hodge (London: Routledge, 1990). In the latter, I note especially, articles on "Realism" by W. H. Newton-Smith; "The Development of the Philosophy of Science 1600-1900" by Ernan McMullin; and "The Development of the Philosophy of Science since 1900" by M. J. S. Hodge and G. N. Cantor. Articles in the former volume that treat such issues are too numerous to list but include: "Confirmation, Semantics, and the Interpretation of Scientific Theories" by Richard Boyd and "The Confutation of Convergent Realism" by Larry Laudan [3] See Virginia P. Dawson, "The Limits of Observation and the Hypotheses of Georges Louis Buffon and Charles Bonnet" in Beyond the History of Science, ed. Elizabeth Garber (Bethlehem, PA: Lehigh University Press, 1990), pp. 107-125. A crucial text is Diderot's Pensées sur l'Interpretation de la Nature (1754); see the helpful introduction to this work in Laurent Versini, ed. Diderot Oeuvres, 5 vols (Paris: Robert Laffont, 1993), 1.555-57. Parts of this and other scientific works of Diderot are translated in Jean Stewart and Jonathan Kemp, Diderot Interpreter of Nature (New York: International Publishers, 1937). Les Bijoux Indiscrets appears in volume two of Versini, Diderot Oeuvres; important in this context are the chapters on "The State of the Academy of Science of Banza" and "Mangogul's Dream; or, A Journey to the Land of Hypotheses." The most recent translation is by Sophie Hawkes, The Indiscreet Jewels (New York: Marsilio, 1993), where these appear as Chapters 9 and 32. On Reid, see Larry Laudan, Science and Hypothesis: Historical Essays on Scientific Methodology (Dordrecht, Netherlands: 1981, Chapter 7.
[4] The standard text on this process is Gerd Buchdahl, Metaphysics and the Philosophy of Science, The Classical Origins: Descartes to Kant (Cambridge, MA: The MIT Press, 1969). See also Shapiro, Probability and Certainty and, on the shift in rhetoric from syllogistic to probabilistic argumentation, Wilbur Samuel Howell, Eighteenth-Century British Logic and Rhetoric (Princeton: Princeton University Press, 1971). [5] See Jürgen Habermas, The Structural Transformation of the Public Sphere: An Inquiry into a Category of Bourgeois Society, trans. Thomas Burger (Cambridge, MA: The MIT Press, 1991). For commentary on Habermas's ideas about the "public sphere," see Craig Galhoun, ed. Habermas and the Public Sphere (Cambridge, MA: The MIT Press, 1993). On the public circulation of debates about science and public experiments, see especially, Larry Stewart, The Rise of Public Science: Rhetoric, Technology, and Natural Philosophy in Newtonian Britain, 1660-1750 (Cambridge: Cambridge University Press, 1992. But see also, among other essays in David Goodring, Trevor Pinch, and Simon Schaffer, eds. The Uses of Experiment (Cambridge: Cambridge University Press, 1989), Simon Schaffer's "Glass Works: Newton's Prisms and the Uses of Experiment," pp. 67-104. [6] See Alexander Welsh, Strong Representations: Narrative and Circumstantial Evidence in England (Baltimore: The Johns Hopkins University Press, 1992). In order to avoid repeated notes, I group here some standard texts that have helped to inform me about early modern science and that touch on many issues raised here. Barry Gower, The Scientific Method: An Historical and Philosophical Guide (London: Routledge, 1997). Steven Shapin, The Scientific Revolution (Chicago: University of Chicago Press, 1996). Thomas L. Hankins, Science and the Enlightenment (Cambridge: Cambridge University Press, 1985. A. Rupert Hall, The Revolution in Science, 1500-1750 (London: Longman, 1983) and From Galileo to Newton (New York: Dover, 1981). Ryan D. Tweny, Michael E. Doherty, & Clifford R. Mynatt, On Scientific Thinking (New York: Columbia University Press, 1981). Ralph M. Blake, Curt J. Ducasse, and Edward H. Madden, Theories of Scientific Method: The Renaissance through the Nineteenth Century (Seattle: University of Washington Press, 1960. [7] Stephen Shapin and Simon Schaffer, Leviathan and the Air-Pump: Hobbes, Boyle, and the Experimental Life (Princeton: Princeton University Press, 1985), Chapters 2 and 6. Hunter suggests the relevance of this book to the study of the novel in his article, "Robert Boyle and the Epistemology of the Novel." [8] Johnson's Dictionary of the English Language (1755) is cited from Harrison's edition (London: Harrison, 1786). See also the important anonymous articles in the Encyclopédie on "probability" and "verisimilitude" (both dating from 1765). [9] David Hume, A Treatise of Human Nature, 2nd ed., eds. P. H. Nidditch and . A. Selby-Bigge (Oxford: Oxford University Press, 1978, Appendix, p. 630. Further quotations cited in text.
[10] On this huge issue in Hume of causation and induction, especially his skeptical analysis of the bases of induction, see Buchdahl, Metaphysics and the Philosophy of Science, Chapter 6. See also, John Biro, "Hume's New Science of Mind" and Alexander Rosenberg, "Hume and the Philosophy of Science" in David Fate Norton, ed. The Cambridge Companion to Hume (Cambridge: Cambridge University Press, 1993), pp. 33-64 and 65-89. [11] Letter to William Warburton, 19 April 1748 in John Carroll, ed., Selected Letters of Samuel Richardson (Oxford: Clarendon Press, 1964), p. 85. [12] See Buchdahl, Metaphysics and the Philosophy of Science, Chapter 6. For eighteenth-century summations, see the 1765 article on "induction" in the Encyclopédie; also Johnson's Dictionary, where "induction" is defined as "when, from several particular propositions, we infer one general." He quotes Newton's Opticks: "Although the arguing from experiments and observations by induction be no demonstration of general conclusions, yet it is the best way of arguing which the nature of things admits of, and may be looked upon as so much the stronger by how much the induction is more general; and if no exception occur from phaenomena, the conclusion may be general." For twentieth-century discussions of induction by Karl Popper, Hilary Putnam, and W. V. O. Quine, among others, see also Boyd, et al, The Philosophy of Science. [13] Tom Jones, ed. John Bender and Simon Stern, with Introduction by John Bender (Oxford: Oxford University Press, 1996), Book V, Chapter 4, p. 194. [14] R. F. Brissenden, ed., Clarissa: Preface, Hints of Prefaces, and Postscript (Los Angeles: William Andrews Clark Memorial Library, 1964) p. 5. For discussion of the juridical context of this passage, see Welsh, Strong Representations, pp. 69-70. See also Bender, Imagining the Penitentiary, Chapter 6. [15] See McKeon, The Origins of the English Novel, pp. 25-64. [16] Sir Isaac Newton, Principles of Natural Philosophy and His System of the World, trans. Andrew Motte, ed. and revised Florian Cajori (Berkeley: University of California Press, 1934). The "General Scholium" appears on pp. 543-547. For a convenient account of the meaning of the phrase "hypotheses non fingo" see Derek Gjertsen, The Newton Handbook (London: Routledge & Kegan Paul, 1986), p. 266; Alexandre Koyré persuasively argued that fingo meant "feign" in Newtonian Studies (London: Chapman & Hall, 1965). For an excellent account of background and context of Newton's thought on hypotheses, see Ernan McMullin, "Empiricism and the Scientific Revolution" in Charles S. Singleton, ed. Art, Science, and History in the Renaissance (Baltimore: The Johns Hopkins University Press, 1967). The complex question of the relationship of Newton's scientific work to his religious belief is a major theme in Richard S. Westfall in Never At Rest: A Biography of Isaac Newton (Cambridge: Cambridge University Press, 1980); see also Markley, Fallen Languages. [17] The text of Preface to Robinson Crusoe is modernized from J. Donald Crowley's edition (Oxford: Oxford University Press, 1972; The "Author's Preface" to The Farther Adventures of Robinson Crusoe is cited, with my italics, from J. B. Yeats's edition (London: J. M. Dent, 1899). The most convenient place to find the prefaces and Gildon's pamphlet is in Ioan M. Williams, ed. Novel and Romance, 1700-1800, a Documentary Record (New York: Barnes & Noble, 1970), pp. 56-70. The issue of parable and allegory is central to Defoe's debate with Gildon; interesting, in the "General Scholium" Newton discusses both the heuristic value of allegories describing God's powers and their limitations as fictions when considered as scientific explanations of cause. [18] Newton, Principles of Natural Philosophy and His System of the World, p. 400. The "Rules of Reasoning" are discussed by McMullin in "Empiricism and the Scientific Revolution" and by Westfall in Never At Rest, p. 801. The Principia underwent many revisions, the chief of which are clarified on p. 466 of the article on the book in Gjertsen, The Newton Handbook. As this article points out, Newton retreated from applying the term "hypothesis" to his method, even though he had frequently used hypotheses before. For Florian Cajori's full listing of the chief instances, see Principles of Natural Philosophy and His System of the World, Appendix, pp. 671-76. The interesting notion of "approximate truth" is still alive in the philosophy of science; it arises in a number of articles in Richard Boyd, et al, The Philosophy of Science (see index). [19] Quoted from George A. Aitken, ed., Serious Reflections During the Life and Surprising Adventures of Robinson Crusoe (London: J. M. Dent, 1895), p. 99. Discussed by McKeon, The Origins of the English Novel, pp. 120- 21. Interestingly, the context of Defoe's statement is a long chapter titled, "Of the Immorality of Conversation, and the Vulgar Errors of Behavior," the drift of which is that conversation is inherently corrupting and productive of error and falsification. This is the opposite of the ideal of more modern Enlightenment ideal of conversation as productive of knowledge advanced in Addison's Spectator and paradigmatic for Habermas. [20] Hans Vaihinger, The Philosophy of 'As if': A System of the Theoretical, Practical and Religious Fictions of Mankind, trans. C. K. Ogden (London: Routledge & Kegan Paul), 1935. Vaihinger works through every category of "fiction" in this book (see pp. 15-77 for summary). He defines hypotheses as strictly separate from fictions of all kinds, and especially poetic fictions, because of their verifiability (pp. 81-90). I show how difficult it proved to maintain this distinction in the eighteenth century and I believe that discussions in the philosophy of science show that it remains so now. Vaihinger, a neo-Kantian, makes interesting arguments for the value of fictions in the conduct of practical reason. [21] Wesley Trimpi, "The Ancient Hypothesis of Fiction: An Essay on the Origins of Literary Theory," Traditio 27 (1971), p. 54-55. On the changing idea of "hypothesis" in Copernicus, Boyle, and others, see McMullin, "Empiricism and the Scientific Revolution." [22] Wesley Trimpi, Muses of one Mind: the Literary Analysis of Experience and Its Continuity (Princeton: Princeton University Press, 1983). p. 30, where he quotes Quintilian, 5.10.95-96; see pp. 50- 51 for hypothesis as plot outline and p. 25 for Trimpi's basic definitions. See also, George Kennedy, ed. The Cambridge History of Literary Criticism. Vol. 1, "Classical Criticism" (Cambridge, Cambridge University Press, 1989), pp. 198 and 207 on hypothesis, plus numerous entries on fictionality. [23]Quoted by Simon Schaffer in "Natural Philosophy and Public Spectacle in the Eighteenth Century" History of Science 21 (1983), 1. He cites Joseph Priestley, "Lectures on History and General Policy" in J.T. Rutt, ed., the Theological and Miscellaneous Works of Joseph Priestley, 25 vols (London, 1817-31)24: 27-28. [24] Lettres Philosophiques, ed. Raymond Naves (Paris: Garnier, 1988), p. 232; this quotation is from Voltaire's 1739 revision of the opening of the letter (my translation). The reference to Castel is from a letter of March, 1738, cited in Jean Ehrard, L'Idée de Nature en France, 2 vols. (Paris: S. E. V. P. E. N., 1963), I, 121. I am indebted for J. B. Shank for these references. Many more similar uses of the word "roman" appear in the on-line ARTFL database of French texts of the period maintained at the University of Chicago. [25] The long article on "phenomenon" by D.F. in the Supplement contains a more skeptical view of hypotheses; the article has not been attributed. The standard work attributing articles is the Inventory of Diderot's Encyclopédie by Richard Schwab, Walter E. Rex, and John Lough, which appeared in Theodore Besterman, ed., Studies on Voltaire and the Eighteenth Century, as volumes 80, 83, 85, 91-93 (1971-72). See also, Jacques Proust, Diderot et L'Encyclopédie, 3rd ed. (Paris: A. Michel, 1995) and John Lough, Essays on the Encyclopédie of Diderot and d'Alembert (London: Oxford University Press, 1968). For biographies of the authors, including Marmontel and La Chapelle, see Frank A. Kafker and Serena L. Kafker, The Encyclopedists as Individuals: A Biographical Dictionary of the Authors of the Encyclopédie (Oxford: Voltaire Foundation, 1988). I am grateful to Mark Olsen and Jack Iverson at the ARTFL project for their views on the authorship of the articles cited in this essay. [26] Nelson Goodman, Ways of Worldmaking (Indianapolis, Indiana: Hackett, 1978). [27] See Buchdahl, Metaphysics and the Philosophy of Science and or the article on "Realism" by W. H. Newton-Smith in R. C. Olby, et al, Companion to the History of Modern Science, pp. 181-195; or, for instance, Ian Hacking, Representing and Intervening: Introductory Topics in the Philosophy of Natural Science (Cambridge: Cambridge University Press, 1983). [28] Versini, Diderot Oeuvres, I.563. My translation, with reference to that in Stewart and Kemp, Diderot, Interpreter of Nature, p. 43. [29] See Bender, Imagining the Penitentiary, pp. 36- 37 and 269 n. 49. [30] G. E. du Châtelet-Lomont, Institutions de physique (Paris: Prault fils, 1740), p. 87. My italics; may translation. Cited in Phillip R. Sloan, "L'Hypothéisme de Buffon: sa place dans la philosophie des sciences du dix-huitième siècle," in Buffon 88: Actes du colloque international pour le bicentennaire de la mort de Buffon (Paris: J. Vrin, 1992), p. 212. [31] These questions thread obsessively through the large number of prefaces and pamphlets scout the novel conveniently collected in Williams, Novel and Romance. [32] Catherine Gallagher, Nobody's Story: The Vanishing Acts of Women Writers in the Marketplace, 1670-1820 (Berkeley: University of California Press, 1994), especially Chapter 4. [33] Williams, Novel and Romance, p. 281. [34] Martin C. Battestin, "'Tom Jones': The Argument of Design," pp. 289-319 in Henry Knight Miller, Eric Rothstein, and G. S. Rousseau, eds., The Augustan Milieu, Essays Presented to Louis A. Landa (Oxford: Clarendon Press, 1970), p. 307. [35] Laurence Sterne, The Life and Opinions of Tristram Shandy, Gentleman, ed Ian Watt (Boston: Houghton Mifflin, 1965), pp. 114-16. [36] The quotations come from Horace Walpole, The Castle of Otranto, ed. W. S. Lewis (Oxford: Oxford University Press, 1964), pp. 4, 5, and 7-8. [37] Friedrich A. Kittler, Aufschreibesysteme 1800/1900 (Munchen: Fink, 1985); published in English as Discourse Networks 1800/1900, trans. Michael Metteer and Chris Cullens, with foreword by David E. Wellbery (Stanford: Stanford University Press, 1990). It is worth noting that efflorescence of fiction that Walpole advocates comes in the same year as the publication of Thomas Reid's An Inquiry into the Human Mind on the Principles of Common Sense (1764), though Reid's formal attack on hypotheses is not fully developed until his Essays on the Intellectual Powers of Man (1785). On Reid, see Laudan, Science and Hypothesis, Chapter 7 and Keith Leher, Thomas Reid (London: Routledge, 1989). [38] Welsh, Strong Representations, p. ix-x. [39] Jean-Jacques Rousseau, Discours sur l'origine et les fondements de l'inégalité parmi les hommes; my translation, with reference to Lester G. Crocker, ed., The Social Contract and Discourse on the Origin of Inequality (New York: Pocket Books, 1967), p. 177; and Roger D. and Judith R. Masters, trans., The First and Second Discourses (New York: St. Martin's Press, 1964, p. 103. [40] See Dawson, "The Limits of Observation and the Hypotheses of Georges Louis Buffon and Charles Bonnet." A convenient collection of early reviews of the Histoire naturelle appears in A convenient collection of reviews of the Histoire naturelle is appears in John Lyon and Philip R. Sloan, ed. and trans., From Natural History to the History of Nature: Readings from Buffon and His Critics (Notre Dame: University of Notre Dame Press, 1981), pp. 213-245. [41] Quoted, in my translation, from Histoire et théorie de la terre (1749) by Philip R. Sloan in "L'Hypothéisme de Buffon," p. 218. [42] Buffon's "Initial Discourse" is cited here from Lyon and Sloan, From Natural History to the History of Nature: Readings from Buffon and His Critics, pp. 122-23. [43] I refer to the section of Buffon's Natural History entitled "Of Reproduction in General"; see Selections from Natural History General and Particular (New York: Arno Press, 1977), pp. 17-38, which reproduces pages from the English edition of 1780-1785, printed for W. Creech, Edinburgh. [44] I thank Mary Poovey both for insightful comments that were crucial in the revision of this essay for publication and for summarizing its arguments along the lines given in this paragraph. Her forthcoming book From Rhetoric to Fact: Sciences of Wealth and Society from Double-Entry Bookkeeping to Statistics (University of Chicago Press) is a major statement in the field of this essay's concern. I also benefited from hearing two conference talks on changes in the concept of "fact" by Lorraine Daston (who is preparing a book on the topic), as well as from her comments on an earlier version of this paper. I also wish to thank Michael McKeon, Clifford Siskin, Elizabeth Heckendorn Cook, and, especially, H. U. Gumbrecht for important assistance. [45] William B. Warner suggested the importance of Dunlop's work and its title, as well as offering many other valuable comments on drafts of this essay. The use of the term "fiction" to mean "novel" already had some earlier currency, for instance in the passage from Anna Laetitia Barbauld and her brother cited above. |