[In the excerpt that follows, Shaw argues that a "negative, minimal definition of historical fiction" is necessary in order to accommodate the variety of interpretations regarding the role of history in the historical novel's structure.]

When critics discuss literary groups and genres, they are usually doing more than indulging in the pleasures of the taxonomical imagination. Genres help us sense the lay of the literary land. They imply questions and sometimes answers: we see a forest, or at least clumps of trees, instead of trees. In other parts of life, we constantly make distinctions that are like generic distinctions in literary studies, and they matter. As we know, an attempt to correct social injustice may dictate very different actions depending on the groups it singles out for attention. Debates between microhistorians and macrohistorians hinge on the same problem; they also remind us that the time span we choose to think about has a significant impact on the conclusions we draw about a given topic. It may be true that in the long run, we are all dead, but such a perspective is more useful to a mystic than to a mercenary. When we experience a work of literature, we employ in a refined and complex way our general ability to see the world in terms of significant groups and patterns. Making sense of a work rests upon knowing what to expect from it, understanding how to take it in. This in turn implies that we have a sense of what sort of thing it is, how it works, what its rules are. Beyond that, generic assumptions allow us—indeed, force us—to focus on some things at the expense of others; they can make certain aspects of a text disappear or seem trivial.…

What is the historical novel? In attempting to answer this question, it would seem advisable, before plunging into speculations about historiography and the nature of a truly historical outlook, to ask what sort of term "the historical novel" is in the first place. How does it differ from other groups of novels—the picaresque novel, the industrial novel, the sentimental novel, the eighteenth-century novel? Which is it most like? A simple but accurate answer is that the term historical novel denotes a kind of novel which can be differentiated from other groups of novels not in terms of a defining compositional technique (the picaresque novel), nor through its power to evoke a set of emotions (the gothic or sentimental novel), and certainly not in terms of the period in which it was written (the eighteenth-century novel). Instead, the principle of differentiation involves the milieu represented, which makes the closest parallel in our list the industrial novel. Though it seems fair to say that the industrial novel is a narrower category, it is the same sort of category as the historical novel.

A convenient way of extending this simple, intuitive notion that historical novels are works that in some way represent historical milieux is to speak in terms of fictional probability. A character or incident in a novel can be probable in either or both of two ways. We usually think of probability as involving fidelity to the external world that a work represents. Some eighteenth-century readers of Clarissa found Mrs. Sinclair's house in London improbable because they could not believe that such carefully contrived dens of iniquity actually existed. One might complain that Clarissa herself violates probability in this sense because she is too good to be true—we have never met anyone like her in the world in which we live. In the historical novel, anachronisms and mistakes of historical fact are responsible for breaches of probability in this sense. But probability can also depend upon how consistently a work follows its own internal rules and patterns. Soliloquies in drama are probable in this second sense, but not in the sense of being faithful reproductions of the behavior we expect from our fellows in their everyday affairs. In general, the more stylized a work becomes, the more these two kinds of probability diverge; the more directly mimetic it is, the more they coalesce. Probability involves our sense of a novel's "fit," both the way it fits the world it imitates and the way its parts fit together to produce a unified whole. A novel's power to illuminate life and its intrinsic beauty as a formed work of art depend in large measure on its probability in both senses.

The concept of fictional probability implies a way of defining historical fiction. We can say that while in most novels probability stems from our general ideas about life and society, in historical novels the major source of probability is specifically historical. Though many kinds of novels may incorporate a sense of history, in historical novels history is, as the Russian Formalists would put it, "fore-grounded." When we read historical novels, we take their events, characters, settings, and language to be historical in one or both of two ways. They may represent societies, modes of speech, or events that in very fact existed in the past, in which case their probability points outward from the work to the world it represents; or they may promote some sort of historical effect within the work, such as providing an entry for the reader into the past, in which case the probability points inward, to the design of the work itself. In Waverly, Fergus Maclvor has both internal and external probability, while Edward Waverley's probability rests primarily on the way in which he furthers the novel's historical design. Fergus is a faithful composite picture of the Highland Jacobite nobility, providing a good external portrait of them and also representing the historical weaknesses that in Scott's opinion doomed their movement. Waverly, on the other hand, is the reader's entry into the novel. He functions primarily as a fictional device, allowing the historical import of the novel to be felt with maximum force. The idea of internal probability allows us to see why a work can become more historical, not less historical, if it rearranges individual aspects of the historical record for the sake of demonstrating a larger pattern.

Historical novels, then, are works in which historical probability reaches a certain level of structural prominence. This may seem an impotent and lame conclusion, objectionable on several counts. It is negative and minimal. It is vague in terms of what counts as historical. It creates distinctions of degree, not kind; and in particular, it does not indicate the kind of prominence history must have in the structure of a truly historical novel. In fact, these qualities are virtues. Because the definition is vague in terms of what counts as historical, it leaves open the possibility that history may mean different things in different works. The definition works in terms of differences of degree, not kind, but it should: the modern historical novel arose as part of the rise of historicism, which made a sense of history part of the cultural mainstream and hence available to novels in general, not simply to historical novels. But the definition's greatest strength is that it does not specify what role history must play in a novel's structure if we are to consider that novel a work of historical fiction. One of my main contentions is that we cannot make sense of historical fiction unless we recognize that history plays a number of distinctly different roles in historical novels.… A negative, minimal definition of historical fiction leaves the way clear for these necessary distinctions.

In most respects, historical fiction depends upon the formal techniques and cultural assumptions of the main traditions of the novel. Because of this dependence, it does not have a significant history apart from the history of the novel as a whole. What is often called the classical historical novel begins with Scott; but the important line of fictional development runs not from Scott to the historical novelists who followed him, but instead from Scott to such masters of European fiction as Balzac, Dickens, and even (so argues Louis Maigron [in Le roman historique a l'epoque romantique: Essai sur l'influence de Walter Scott, 1898]) Flaubert. The authors who produce the best historical novels after Scott tend, with the exception of Cooper and Tolstoy, to be masters of other kinds of writing, who enter the field with one or two attempts, as Dickens, Thackeray, and Hugo do. Georg Lukács [in The Historical Novel, 1937] is in my opinion essentially accurate in describing the history of the novel as a great stream from which tributaries branch off, only to rejoin and further enrich it in due course. Scott's works form such a tributary: he branches off from the eighteenth-century novel, discovers in artistic terms the rich significance of history, and then reunites with the mainstream of nineteenth-century fiction through his influence on Balzac, enriching it with new materials, insights, and techniques.

Since they lack a history of their own, the most useful way to group historical novels historically is in terms of coherent movements of the novel as a whole, and of the esthetic and cultural presuppositions that underlie them. The realist novel, which begins with Richardson and finds its greatest achievement in the works of Eliot and Balzac and Tolstoy, is such a movement. I shall call such works "standard" novels; the group of historical novels which derives its unity from its relationship with standard fiction then becomes "standard historical novels.…" They all employ the formal techniques of standard fiction, and in particular, they use the plotted action, which creates in the reader a pattern of hopes and fears for one or more protagonists, as their formal basis. They also share with the standard novel a set of broad cultural assumptions that provide the grounds for their intelligibility and are the ultimate source of their "realism." The situation of historical fiction in our own century becomes more complex. As the novel in general changes, new forms of historical fiction emerge. But strong continuities with nineteenth-century forms also persist in such distinguished historical novelists as Marguerite Yourcenar or H. F. M. Prescott.

I have suggested that no single quality of historical insight defines historical fiction. But since we have narrowed our sights to the standard historical novel, can't we say something more specific about the kind or kinds of historical vision they embody? We can indeed, but only within limits. The historian Herbert Butterfield and the literary critic Avrom Fleishman have both tried in different ways to define the quality historical novels share, and their discussions are useful here. Both define historical fiction by differentiating it from historiography. For Butterfield [in The Historical Novel, 1924], historiography attempts to "make a generalisation, to find a formula," because it views history as "the whole process of development that leads up to the present." The historical novel, by contrast, attempts to "reconstruct a world, to particularise, to catch a glimpse of human nature." The task of the historical novelist is to render the unique "atmosphere" of an age in the past, to "recapture the fleeting moment." For Fleishman [in The English Historical Novel: Walter Scott to Virginia Woolf, 1971], by contrast, the historical novelist accomplishes something more like the task Butterfield sets the historian: "What makes a historical novel historical is the active presence of a concept of history as a shaping force." Both critics are clearly drawing on the achievements of historicism for their definitions, a procedure that seems entirely in order since the rise of the historical novel is bound up with the rise of historicism in general. But Butterfield invokes what we might call a minimal historicist vision; Fleishman, historicism at its most powerful and dignified.

Identifying the historicity of historical novels with "the active presence of a concept of history as a shaping force" seems to have much to recommend it. Probably the most important aspect of the historicist view of the past is its recognition that history shapes human beings through specific and unique social mediations. This need not imply the view often attributed to Hegel—the idea that history is a vast teleological progression leading relentlessly toward one divine event—though such a belief is one of its extreme potentialities. It does involve a sociological sense of both past and present, a recognition that societies are interrelated systems which change through time and that individuals are profoundly affected by their places within those systems. The greatest modern critics of nineteenth-century fiction from a historical point of view, whatever their other differences, agree that the creation of this grasp of social-historical milieux is its principal achievement. For Erich Auerbach, it is the prerequisite for a fully serious treatment of everyday life; for Lukács, it involves a fundamental discovery concerning the meaning of historical process itself. Both critics describe this development as the discovery of "the present as history," a phrase which derives ultimately from Hegel. In asserting that historical novels are defined by their sense of "history as a shaping force," Fleishman would thus appear to have rendered them an important service: he has acquitted them of the charge of portraying mere local color, finding in them instead the historical insights we associate with historicism at its most developed.

But Butterfield, writing at a time when valuing historical particularity caused critics less embarrassment than it does now, is closer to the truth about the historical probability that serves as a basis for the standard historical novel. The problem with Fleishman's mode of definition is that it saves too much too quickly, giving historical fiction a cognitive dignity that is unearned. The works of Harrison Ainsworth betray no insight whatever into "history as a shaping force," but we unhesitatingly call them historical novels. Fleishman's discussion quickly slides from defining historical fiction to finding a criterion for "authentic" historical fiction, a separate issue for which his maximal kind of definition is entirely appropriate.

The recognition that human beings are part of a larger historical process is not the source of the distinctively historical probability that distinguishes standard historical fiction, though the best historical novels certainly convey it. The probability that distinguishes standard historical novels rests on a simpler notion—the realization that history is comprised of ages and societies that are significantly different from our own. We can call this idea the recognition of "the past as past" (Ainsworth's fiction doesn't really measure up to this criterion either, but it is at least possible to recognize in his sensationalistic use of historical atmosphere a debased version of it, whereas any connection whatever with a notion of historical process or the present as history is in his case unimaginable.) The recognition of the past as past can lead to a sense of history as a process, and perhaps it ought to, but in fact it has not always done so in historical fiction.

By arguing that historical fiction need not view history as a process, I am parting company not only with Fleishman but with Lukács as well, though in a different way. For Lukács, the historical novel arises in the works of Scott when Scott discovers on an esthetic level that history is a process in which the past acts as the necessary precondition for the present. Unlike Butterfield or Fleishman, however, Lukács simply isn't interested in the problem of defining historical fiction, and given his distrust of "mere formalism," it is hard to see how he could be. He pursues instead the question of how the historical spirit comes to consciousness in literature. With more consistency than other writers who hold similar views concerning what is truly historical about historical novels, Lukács believes that historical fiction does not constitute a genre separate from the European realist novel as a whole. In his version of Marxist esthetics, a truly separate genre can arise only from a new vision of reality, and the truly historical novel shares (and in fact helped to create) the vision of reality we find in genuinely realistic novels. Historical fiction is thus part of a larger fictional genre, realist fiction, which is characterized by the mode of knowledge it embodies. This mode of knowledge provides, in the process he calls "preparatory esthetic processing," a necessary but not sufficient condition for the creation of realist fiction.

My discussions of Lukács, Butterfield, and Fleishman have revealed a number of fundamental differences among them, but they are united in believing historical fiction to be fundamentally a mode of knowledge. Such a view has its attractions. Who would want to deny that the best historical fiction can add to the richness of our sense of history, or that the structuring of history in great historical fiction may have cognitive value? Narrativist philosophers of history have recently argued with some persuasiveness that historical understanding itself may proceed according to the logic of narrative discourse, not of science. But it seems important for a number of reasons to oppose the idea that historical novels, or even standard historical novels, embody a defining vision of history in more than a minimal way.

Such an idea can become quickly and narrowly prescriptive in practice, blinding us to the workings of novels that embody a vision of history we do not respect. It is tempting to say that works which embody a historical vision we find uninteresting or unacceptable, or that seem to have no historical vision at all, are not "really" historical novels. But it is more useful to discriminate between great and mediocre historical novels than to exclude imperfect works from the group—a procedure that logically tends to produce a group containing one and only one true member. A different consequence appears in Fleishman's criticism. As we have seen, he believes that historical novels are characterized by "the active presence of a concept of history as a shaping force." He is also interested in tracing the history of historical fiction in England, which of course implies that there is a significant history to trace. Each of these premises raises problems. Combined, they cause him to exaggerate and homogenize the level of historical consciousness in the works he considers. Only by doing so can he produce a developing tradition of English historical fiction.

The search for a specific way of perceiving history which defines historical fiction is in my view a local manifestation of the understandable but unacceptable desire to separate literary discourse from "scientific discourse" and thus to save for literature its own distinct significance. This view, which places great stress on differentiating between true and false poetic modes, is most familiar as one of the cornerstones of the New Criticism as practiced by Cleanth Brooks and others. Lukács from his Marxist point of view has come up with a similar if somewhat more elegant procedure involving three levels of discourse—everyday speech, scientific discourse, and (mediating between the two on a whole series of intermediate levels) the language of literature. This is not the place to weigh the merits of such views extensively. I shall merely say that in my view "scientific discourse" in the sense required by Brooks and the others has been shown not to exist; their argument is based upon a false dichotomy. Literary works in general do not embody a distinct mode of knowledge, though they certainly can impart knowledge. Though literary works can have cognitive claims, they are in the first instance verbal constructions designed to create certain effects through the disposition of their parts.

For our present purposes, the idea that historical fiction is a mode of telling the truth about history is objectionable chiefly because it does not account for the very different formal status that visions of history have in fact assumed in historical fiction. Such a definition excludes works we all call historical novels. Furthermore, if we adopt such a definition, we must conclude that most great nineteenthcentury novels are historical novels, which renders the concept "historical novel" useless as a conceptual aid and falsifies the strong intuitive impression that leads readers to give the group a name in the first place. In practice, such an emphasis also tends to exclude or preclude problems of artistic form and effect, operating as if historical novels conveyed unmediated historical doctrine. Finally, the idea that historical fiction is a mode of historical knowledge leaves as a complete mystery what is surely the most striking fact about these works.