[48]
HISTORICISM, HISTORY, AND THE FIGURATIVE IMAGINATION
HAYDEN V. WHITE Discussions of "historicism" sometimes proceed on the assumption that it consists of a discernible and unjustifiable distortion of a properly "historical" way of representing reality. Thus, for example, there are those who speak of the particularizing interest of the "historian" as against the generalizing interests of the "historicist." Again, the "historian" is supposed to be interested in elaborating points of view rather than in constructing theories, as the "historicist" wishes to do. Next, the "historian" is supposed to favor a narrativist, the "historicist" an analytical mode of representation. And finally, while the "historian" studies the past for its own sake or, as the phrase has it, "for itself alone," the "historicist" wants to use his knowledge of the past to illuminate the problems of his present or, worse, to predict the path of history's future development.1[49]
As can readily be seen, these characterizations of the differences between a properly "historical" and a "historicist" approach to history correspond to those that are conventionally used to differentiate "historiography" from "philosophy of history." I have argued elsewhere that the conventional distinctions between "historiography" and "philosophy of history" obscure more than they illuminate of the true nature of historical representation.2 In this essay I will argue that the conventional distinctions between "history" and "histori-cism" are virtually worthless. I will suggest, on the contrary, that every "historical" representation however particularizing, narrativist, self-consciously perspectival, and fixated on its subject matter "for its own sake" contains most of the elements of what conventional theory calls "historicism." The historian shapes his materials, if not in accordance with what Popper calls (and criticizes as) a "framework of preconceived ideas,"3 then in response to the imperatives of narrative discourse in general. These imperatives are rhetorical in nature. In what follows I shall seek to show that in the very language that the historian uses to describe his object of study, prior to any effort he may make formally to explain or interpret it, he subjects that object of study to the kind of distortion that "historicists" impose upon their materials in a more explicit and formal way.To raise the question of the rhetoric of historical discourse is to raise the problem of the nature of description and analysis in fields of study which, like historiography, have not yet attained to the status of sciences in the way that physics, chemistry, and biology have done. I leave aside for the moment the point made by Claude Levi-Strauss to the effect that history has no method uniquely its own, nor indeed any unique subject matter; and that its fundamental technique, which consists of the arrangement of the events it would analyze in the serial order of their original occurrence, is simply a
[50]
preliminary phase of any analysis worthy of the designation "scientific."4 But I want to dwell momentarily on Levi-Strauss' contention that in history, as in any field of occurrence that we would submit to analysis, there is a paradoxical relationship between the amount of information that may be conveyed in any given account of that field and the kind of comprehension that we can have of it.Levi-Strauss suggests that "the historical field," the general object of the historian's interest, consists of a field of events which dissolves, at the micro-level, into a congeries of physico-chemical impulses and, at the macro-level into the tidal rhythms of the rise and fall of whole civilizations. In his schema, the micro and macro levels correspond to the limits of a set of explanatory strategies which range from the mere chronicling of particular events, on the one side, to the appeal to comprehensive cosmologies, on the other. The relation between the micro and the macro levels he characterizes in terms of a dyad: information-comprehension. And he states the relation between them in the form of a paradox: the more information we seek to register about any given field of occurrence, the less comprehension we can provide for that field; and the more comprehension we claim to offer of it, the less the information covered by the generalizations intended to explain it.8
It is obvious that here Levi-Strauss has extended to theory of knowledge his own version of the Structuralist concept of the bi-polarity of language: his information-explanation dyad corresponds to the terms used by Roman Jakobson and others to characterize the two axes of language, metonymic and metaphoric poles respectively.6 These two poles of language-use are identified with the axes of combination and selection of any meaningful speech act. This provides the basis for Levi-Strauss' characterization of the relationship between the syntagmatic and the paradigmatic axes of any discourse meant to represent a field of happening having the aspects of both process and structure, diachronicity and synchronicity simultaneously. Thus, at the lower (or micro) limit of the historical field, there is no similarity, only contiguity; at the upper (or macro) limit, there is no difference, only similarity. And so too in the discourse that we would construct in order to represent what we perceive to have happened in "the historical field": the historical discourse seeks to represent the unfolding along a temporal line of a structure whose parts are always something less than the totality which they comprise, and the totality of which is always something more than the sum of the parts or phases which make it up.
I do not wish to dwell on this extension of the theory of language to the
[51]
theory of knowledge at this point. For the moment, I wish merely to note that for Levi-Strauss, all sciences (including the physical sciences) are constituted by arbitrary delineations of the domains that they will occupy between the poles of mythic comprehensions of the totality of experience, on the one side, and the "blooming, buzzing confusion" of individual perceptions, on the other. And in his view, this is especially true of a field such as historiography, which seeks to occupy a specifically human domain that is the supposed "middle ground" between the extremes. But this putative "middle ground" does not emerge apodictically from the welter of events and information that we have of the human past and present; it must be constituted. And it is constituted, Levi-Strauss thinks, by virtue of a conceptual strategy that is mythic in nature, and which identifies the "historical" with the experiences, modes of thought, and praxis peculiar to modern Western civilization. Levi-Strauss asserts that the presumed "coherency" of history, which Western historical thought takes as its object of study, is the coherency of myth. And this is as true of "proper" or conventional narrative historiography as it is of historiography's more highly schematized counterparts in philosophy of history.7Now, by the coherency of myth, Levi-Strauss appears to mean the result of the application of narrative strategies by which basic story units (or clusters of events) are arranged so as to give to some purely human structure or process the aspect of cosmic (or natural) necessity, adequacy, or inevitability. Stories of the foundings of cities or states, of the origin of class differences and privileges, of fundamental social transformations by revolution and reform, of specific social responses to natural disasters, and so on all such stories, he suggests, whether presented under the aspect either of social science or of history, partake of the mythical inasmuch as they "cosmologize" or "naturalize" what are in reality nothing but human constructions which might well be other than what they happen to be. As thus envisaged, to historicize any structure, to write its history, is to mythologize it: either in order to effect its transformation by showing how "unnatural" it is (as with Marx and late capitalism), or in order to reinforce its authority by showing how consonant it is with its context, how adequately it conforms to "the order of things" (as with Ranke and Restoration society). History, Levi-Strauss insists, is never only history of, it is always also history for. And it is not only history for in the sense of being written with some ideological aim in view, but also history for in the sense of being written for a specific social group or public. More: this purpose and direction of historical representation are indicated in the very language which the historian uses to Char-
[52]
acterize his data prior to any formal technique of analysis or explication that he may bring to bear upon them so as to disclose what they "really are" or what they "truly mean."8All this is, one could argue, familiar enough. We have encountered versions of it already in Mannheim's analysis of the ideological bases of the forms of historical consciousness in Ideology and Utopia, and especially in his analysis of the conservative implications of Rankean historicism in his essay "Conservative Thought."9 But Levi-Strauss' treatment of the relation between historical thought and the mythic imagination is much more radical than anything conceived by Mannheim. For Levi-Strauss locates the impulse to mythologize, not in the real or imagined interests of the social groups for which different kinds of historiography might be written, as Mannheim does, but rather in the very nature of language itself. More specifically, he locates the impulse to mythologize in a poetic faculty which shows itself as readily in such putatively "realistic" forms of prose discourse as historiography as it does in the manifestly figurative nature of that form of discourse which is called "poetry" by "civilized" man.
The conflation of the concepts of poetry and prose within a general theory of language as discourse is one of the principal achievements of modern linguistic theory. The implications of this conflation have been especially fruitful for the field of stylistics. As elaborated by Jakobson, the problem of style recalls us to the recognition that every discourse is a mediation between the metaphorical and the metonymic poles of language behavior through the instrumentality of those "figures of speech" originally studied by the classical rhetoricians. In Jakobson's view, stylistics must seek to analyze the poetic dimension of every merely putatively prose discourse, just as it must seek to uncover the prosaic kernel of "message" contained in every manifestly poetic utterance.10 This conflation of the prosaic and the poetic within a general theory of discourse has important implications for our understanding of what is involved in those fields of study which, like historiography, seek to be "objective" and "realistic" in their representations of the world but which, by virtue of the unacknowledged poetic element in their discourse, hide their own "subjectivity" and "culture-boundedness" from themselves.
If Jakobson is right, then historical writing must be analyzed primarily as a kind of prose discourse before its claims to objectivity and truthfulness can be tested. This means subjecting any historical discourse to a rhetorical
[53]
analysis, so as to disclose the poetical understructure of what is meant to pass for a modest prose representation of reality. Such an analysis would provide us, I maintain, with a means of classifying the different types of historical discourse in terms of the modalities of figurative language-use which are favored in them. It would permit us to transcend the analytically worthless classification of historical tracts into two mutually exclusive classes defined by their interest in the particular versus the general, the past versus the present and future, point of view versus theory, and so on; to collapse what is a false distinction between a properly "historical" and a merely "his-toricist" account of history; and to disclose the extent to which a given historical discourse is more accurately classified by the language used to describe its object of study than by any formal analytical techniques it applies to that object in order to "explain" it.A rhetorical analysis of historical discourse would recognize that every history worthy of the name contains not only a certain amount of information and an explanation (or interpretation) of what this information "means," but also a more or less overt message about the attitude the reader should assume before both the data reported and their formal interpretation.11 This message is contained in the figurative elements appearing in the discourse which serve as subliminally projected clues to the reader about the quality of the subject under study. And these figurative elements play a correspondingly more important role as components of the message of the historical discourse precisely in the degree to which the discourse itself is cast in ordinary, rather than technical, language. As thus envisaged, those historians who pride themselves on avoiding the use of all jargon and technical terminology in their descriptions and analyses of their subjects should not be regarded as having avoided falling into "historicism" as a result, but rather as being historicists of a particular kind. I would call them "figurative his-toricists" inasmuch as they remain unaware of the extent to which what they say about their subjects is inextricably bound up, if not identical, with how they say it.
There is, of course, no escaping the determinative power of figurative language use. Figures of speech are the very marrow of the historian's individual style. Remove them from his discourse, and you destroy much of its impact as an "explanation" in the form of an "idiographic" description. But the study of the figurative element in a given historical discourse permits us to characterize the instrumental, pragmatic, or conative dimensions of it. The theory of figures of speech permits us to track the historian in his encodation of a field of happening in what may appear to be only an original and value-free description, but which in reality is a prefiguration of the field that prepares
[54]
us for the formal explanation or interpretation of it that he will subsequently offer.12 As thus conceived, the clue to the "meaning" of a given historical discourse is contained as much in the rhetoric of the description of the field as it is in the logic of whatever argument may be offered as its explanation. If anything, this rhetorical element is even more important than the logical one for comprehending what goes on in the composition of an historical discourse. For it is by figuration that the historian virtually constitutes the subject of the discourse; his explanation is little more than a formalized projection of qualities assigned to the subject in his original figuration of it.All this is highly abstract, of course, and in order to be made convincing requires both theoretical amplification and exemplification. In what follows, therefore, I will try to characterize the historical discourse in somewhat more formal terms and then analyze a passage of "proper" historical prose in order to explicate the relationship that obtains between its manifest and latent (figurative) meanings. After that, I will return to the problem of the relation between "proper" historiography and its historicist counterpart, on the one side, and to some general remarks on the possible types or modes of historical representation suggested by figurative analysis, on the other.
I have argued elsewhere that an historical discourse should not be regarded as a mirror image of the set of events that it claims simply to describe.18 On the contrary, the historical discourse should be viewed as a sign-system which points in two directions simultaneously: first, toward the set of events it purports to describe and, second, toward the generic story form, to which it tacitly likens the set in order to disclose its formal coherence considered as either a structure or a process. Thus, for example, a given set of events, arranged more or less chronologically, but encoded so as to appear as phases of a process with a discernible beginning, middle, and end, may be emplot-ted as a Romance, Comedy, Tragedy, Epic, or what have you, depending upon the valences assigned to different events in the series as elements of recognizable archetypal story-forms.
In this work of emplotment we can perceive the operation of those processes which Freud, in his Interpretation of Dreams, convincingly identifies as components of any poetic activity, whether of the waking or of the sleeping consciousness.14 The historian like any writer of a prose discourse fashions his materials. He may fashion them so as to make them conform to a "framework of preconceived ideas" of the sort that Popper ascribes to Hegel and Marx, or he may fashion them so as to make them conform to a
[55]
"preconceived selective point of view" of the sort that the novelist occupies in his function as the narrator of a story.15 But in either case, his account of the phenomena under consideration will unfold on at least two levels of meaning, which we may liken to the manifest and latent levels of a dream or to the literal and figurative levels of imaginative literature in general.In most discussions of historical discourse, the two levels conventionally distinguished are those of the facts (data or information) on the one side and the interpretation (explanation or story told about the facts) on the other. What this conventional distinction obscures is the difficulty of discriminating within the discourse between these two levels. It is not the case that a fact is one thing and its interpretation another. The fact is presented where and how it is in the discourse in order to sanction the interpretation to which it is meant to contribute. And the interpretation derives its force of plausibility from the arrangement of the facts in the order and manner in which they are presented in the discourse. The discourse itself is the actual combination of facts and meaning which gives to it the aspect of a specific structure of meaning that permits us to identify it as a product of one kind of historical consciousness rather than another.
There should be nothing very surprising about this last contention, since it is a commonplace of historical theory that all historical accounts are "artistic" in some way. I propose to suggest, however, that the artistic component in historical discourse can be disclosed by an analysis that is specifically rhetorical in nature. Moreover, I shall argue, the principal types of historical discourse can be identified with the types of prose discourse analyzed in rhetorical theory in terms of the modes of figurative language-use that they variously favor.
In order to indicate what I have in mind, I will analyze a passage written by a modern "historian" whom no one, I think, would seriously consider to be a "historicist." What I propose to show is that the explanatory effect of this representation of a set of events derives primarily from its appeal to certain conventions of literary characterization which make up the figurative level of the discourse. I will show also that this latent meaning of the discourse can be identified with the very language used to describe the events analyzed. This language-use serves as a "code" by which the reader is invited to assume a certain "attitude" toward the "facts" and the "interpretation" of them offered on the manifest level of the discourse. Here is the passage:
The Republic created by the Constituent Assembly at Weimar lasted in theory for fourteen years, from 1919 to 1933. Its real life was shorter. Its first four years were consumed in the political and economic confusion which followed the Four Years' War; in its last three years there was a temporary dictatorship, half-cloaked in legality, which reduced the republic to a sham long before it was
[56]
openly overthrown. Only for six years did Germany lead a life ostensibly democratic, ostensibly pacific; but in the eyes of many foreign observers these six years appeared as normal, the "true" Germany, from which the preceding centuries and the subsequent decade of German history were an aberration. A deeper investigation might have found for these six years other causes than the beauty of the German character.16I chose this passage "at random" in the sense that I simply opened an anthology of historical writing on the Third Reich and examined a number of synoptic characterizations of the "era" written by historians of different methodological persuasion and ideological convictions. It has the advantage, for my purposes, of being written in standard ordinary English, rather than in technical jargon, and it is manifestly "literary" in style.
The historian who wrote this passage is widely praised as a writer; he is also widely recognized as a no-nonsense purveyor of facts, and as a polemicist of exceptional, though by no means perverse, talents. Moreover, if it were suggested to him that what he has to say, that is, his presentation of the facts and the arguments he offers in support of his explanation of the facts, was indistinguishable from the way he says it, he would in all probability regard this as an affront to his professional competence. But his account of this period of German history is little more than a discourse in which he progressively earns the right to the rhetorical characterization of the events he purports merely to describe and objectively analyze. Like all historical representations, this one too is a progressive encodation, on a deep or figurative level, of events which exist on the surface level as simple description and analysis.
Now, much of this is obvious from the diction of the passage. This alone signals the ironic posture of the writer, not only with respect to the unnamed "foreign observers" of the six years which appeared to them as "normal," but also with respect to the "Germany" of the period as well. But there is more here than simply an ironic tone. Fact and figurative characterization are combined to create an image of an object the real referent of the discourse that is quite different from the manifest referent, namely, Germany itself. This latent referent is constituted by the rhetorical techniques of figuration that are identifiable on the surface of the discourse.
Let us consider, first of all, the factual information contained in the passage. We are told that:
[57]
The rest of the passage consists of innuendo and thinly masked slurs on the naivete of certain "foreign observers" as well as an allusion to a "deeper investigation" that "might have found" for the six years in question "other causes" than the "beauty of German character" and, presumably, a way of penetrating through the "ostensible" form of German history of this period to its obviously corrupt substance.
Now, this passage is a good example, in microcosm, as it were, of the essential elements of any historical discourse. We have, on the manifest level, the chronicle of events which provide the elements of a story with a discernible beginning (1919-23), middle (1923-29), and end (1929-32). This story, in turn, has an identifiable plot-structure which unites these phases into a process describing the unfolding of a pseudo-tragedy. The plot-structure serves as a kind of secondary elaboration of the events making up the chronicle and their arrangement in a story, by disclosing the latent meaning of the manifest representation of the facts. It plays upon our presumed, but formally uninvoked, capacities to "track" the events related in the story and subconsciously to "decode" their subliminally encoded structure as a particular kind of story (a pseudo or satirical tragedy). In other words, the events in the story are encoded by the use of the figurative language in which they are characterized, in order to permit their identification as elements of the particular story-type to which this story belongs.
The provision of this secondary meaning is signalled in the opening two sentences of the passage, in which the duration of the Republic for fourteen years "in theory" is contrasted with its "real life" of only six. This contrast between the theoretical and real "lives" of the Republic quickly moves the subject of the discourse into that category of grotesques which we are accustomed to meet in satire. The principal active verbs used in the exposition: "created," "consumed," "reduced," and "overthrown" themselves serve to
[58] characterize the phases of the archetypal literary fiction to which the Republic's life is being tacitly likened, namely, the pseudo-tragedy.That the Republic's short life was only a pseudo-tragedy is indicated by the fact that the account of the progressive destruction of the Republic is unrelieved by any suggestion of countervailing tendencies in it. Insofar as such tendencies are indicated, they are labelled as only "ostensible." And those "foreign observers" who viewed the middle "six years" as representing the "normal" and "true" Germany are themselves characterized, by a conventional metonymy of "eyes" for "minds," as being as superficial in their perceptions as any mere eye undirected by intelligence might be expected to be. The "deeper investigation" which these "foreign observers" failed to undertake (and which the author presumably has undertaken) alludes by ironic indirection ("might have found") to the "ugliness" of the German character signalled figuratively (that is, ironically) in the reference to its merely seeming "beauty."
Now, I have belabored this really rather innocuous passage of historical prose, which A. J. P. Taylor must have written quite fluently and naturally, to make a simple point. The point is this: even in the simplest prose discourse, and even in one in which the object of representation is intended to be nothing but fact, the use of language itself projects a level of secondary meaning below or behind the phenomena being "described." This secondary meaning exists quite apart from both the "facts" themselves and any explicit argument that might be offered in the extra-descriptive, more purely analytical or interpretative, level of the text. This figurative level is produced by a constructive process, poetic in nature, which prepares the reader of the text more or less subconsciously to receive both the description of the facts and their explanation as plausible, on the one side, and as adequate to one another, on the other.
As thus envisaged, the historical discourse can be broken down into two levels of meaning. The facts and their formal explanation or interpretation appear as the manifest or literal "surface" of the discourse, while the figurative language used to characterize the facts points to a deep-structural meaning. This latent meaning of an historical discourse consists of the generic story-type of which the facts themselves, arranged in a specific order and endowed with different weights, are the manifest form. We understand the specific story being told about the facts when we identify the generic story-type of which the particular story is an instantiation.
This conception of the historical discourse permits us to consider the specific story as an image of the events about -which the story is told, while the generic story-type serves as a conceptual model to which the events are to be likened in order to permit their encodation as elements of a recognizable structure.
[59] A conceptual model may be invoked more or less explicitly and may be presented more or less formally in the effort to explain or interpret the events represented in the narrative. But such formal and explicit invocations of a conceptual model, as in a nomological-deductive argument for example, must be distinguished from the figurative meaning of the historical discourse. The figurative meaning is implicit even in the simple description of the events prior to their analysis, as well as in the story told about them. The story transforms the events from the meaninglessness of their serial arrangement in a chronicle into a hypotactically arranged structure of occurrences about which meaningful questions (what, where, when, how, and why) can be asked. This story element in the historical discourse exists even in the most severe examples of structuralist, synchronic, statistical, or cross-sectional historical writing. Such historical discourse would have no problematic if it did not tacitly distinguish between the serial order of events and some kind of transformation of that order into a structure of which meaningful questions can be asked.It is commonplace, of course, that an historical discourse does not represent a perfect equivalent of the phenomenal field it purports to describe, in size, scope, or the order of seriality in which the events occurred. But this fact is usually construed as a simple reduction by selection, rather than as the distortion which it truly is. The most obvious manner of distortion is the departure from the chronological order of the events' original occurrence, so as to disclose their "true" or "latent" meanings. Here, of course, we must confront the conventional, but never fully analyzed, distinction between the "mere" chronicle and the history properly so-called. Everyone grants that the historian must go beyond the serial organization of events to the determination of their coherency as a structure and assign different functional values to the individual events and the classes of events to which they appear to belong. This task is usually conceived, however, as that of "finding" the story or stories that are supposed to lie embedded within the welter of facts reported in the record or the diachronic series of events as arranged in the chronicle. Actually, however, nothing could be further from the truth of the matter.
No given set of events figures forth apodictically the kind of meanings with which stories provide them. This is as true of sets of events on the scale of an individual life as it is of those spanning a century in a nation's evolution. No one and nothing lives a story. And sequences of events can take on the aspects of a Romance, a Tragedy, or a Comedy indifferently, depending on the point of view from which they are apprehended and the generic story-form chosen by the historian to guide the articulation of the story.
What is involved here is a fashioning of a framework within which to place
[60] events, of different magnitude and complexity, in order to permit their en-codation as elements of different story-types. This fashioning process need not be it stressed entail violations of the so-called "rules of evidence" or the criteria of "factual accuracy" resulting from simple ignorance of the record or the misinformation that might be contained in it. But this fashioning is a distortion of the whole factual field of which the discourse purports to be a representation as is the case in all model-building.This distortion, in turn, may be of two sorts: negative, consisting of the exclusion of facts that might have been included in the representation of the field; and positive, consisting of the arrangement of events in an order different from the chronological order of their original occurrence, so as to endow them with different functions in an integrated pattern of meaning. And here the historian, in his capacity as literary artist, utilizes the same techniques of condensation, displacement, considerations of representability, and secondary elaboration identified by Freud as the psychological strategies used in the "dreamwork" to mask the latent (and real) meaning of a dream behind the manifest or literal level of the dream report.17
To be sure, in the historian's "knowledge-work" these techniques are used to effect an opposite movement, that is, from the welter of facts which have the meaningless structure of mere seriality to the disclosure of their putatively "true" or "real" significance as elements of a comprehensible process. It does not matter whether the techniques of this fashioning process consist of the application of chi-square functions to what appears to be a jumble of random events, or the application of rules of emplotment so as to disclose the "drama" in what appears to be a chaos of happening. The result is a distortion of the whole factual field considered as a totality of all of the events that might be perceived to have occurred within its confines. This distortion may seem more comprehensible than the field in its unprocessed state or when processed only into chronicle. But it is only more comprehensible by reference to the conceptual model which sanctioned its distortion in this way and not another. It is in response to this presupposed conceptual model that the historian "condenses" his materials (that is, includes some events and excludes others); "displaces" some facts to the periphery or background and moves others closer to the center; encodes some as causes and others as effects, joins some and disjoins others in order to "represent" his distortion as a plausible one; and creates another discourse, a "secondary elaboration" running alongside the more obviously representational level of the discourse, which appears usually as a direct address to the reader and provides the ex-
[61]
plicit cognitive grounds (the "rationalization") for the manifest form of the discourse in general.In order to clarify what is involved here, let us return to the passage taken from Taylor's book on Germany. In this brief synoptic characterization of the period between 1919 and 1933, the evidences of condensation are obvious. It does not matter that we have taken only a paragraph, rather than a chapter or a larger part of the text, for analysis. The entire book necessarily condenses its material, not only in the sense of reducing the range of possible representation, that is, the size of the object dealt with, but also in the sense of over-determining certain elements of the object, so as to reveal the pseudo-tragical nature of the totality of events depicted, considered as a completed process. As for the evidence of displacement, this is equally obvious in the juxtaposition of the "real life" of the Republic to its apparent ("theoretical") life. This "real life" is the center of the discourse, while the "theoretical life" is progressively remanded to the periphery through its revelation as an illusion. And so too for the considerations of representability. Evidence of these appear on the surface of the text as a citation of the causes of the fall of the Republic: the "confusion" of the postwar period, the creation of a "temporary dictatorship" which further undermined the spirit (though not the letter) of the Constitution, and the generally "sham" nature of the Republic's political structure. But the real causes of the Republic's failure are indicated only figuratively, as residing in that "ugliness" of the German character which the notion of its only apparent "beauty" ironically invokes.
The two causal statements made in the passage bear further analysis. The first states that the "first four years [of the Republic] were consumed in the political and economic confusion which followed the Four Years' War . . ." Literally, the statement suggests that "confusion" causes political weakness;
but it actually says that "years" can be "consumed" in "confusion." Here the word "years" is a metonymy for "life" which is itself a metaphor for "energies." But the use of a passive verb ("were consumed") further suggests that these "energies" were weak from the first. A similar contrast between what is literally said and what is suggested by figurative inversions can be seen in the second causal statement in the passage: "the temporary dictatorship, half-cloaked in legality, reduced the republic to a sham long before it was openly overthrown." Literally, the statement asserts that the "temporary dictatorship, half-cloaked in legality" (itself a metaphorical characterization suggestive of "sinister forces" at work on the scene) "reduced" the Republic to a "sham." But here the verb used ("reduced") is active, rather than passive, and thereby suggests the power and strength of the Republic's enemies in contrast to the weakness of its supporters. This implied contrast permits the reader to accept the explanation of the Republic's fall
[62] that will ultimately be provided as a plausible one. After all, it is hardly surprising that strong and active powers should succeed in destroying weak and confused ones. It is in the nature of things that "real" entities overcome "sham" ones.We are now in a position to identify the controlling metaphor of the whole passage, that which mediates between the literal and the figurative dimensions of the discourse, as disclosed in the word "sham." This word is etymologically associated with English "shame" and in its earliest recorded uses in English connotes "trick," "fraud," and "counterfeit."18 It is this metaphor, with its suggestions of bad faith, hollowness, and mere appearances, that sanctions the use of the verbs demarcating the successive stages of the process of disintegration in the Republic's "life": "created," "consumed," "reduced," and, finally, "overthrown." It is also this metaphor that sanctions Taylor's ironic attitude vis-a-vis both the subject of his discourse, the Weimar Republic, and those "foreign observers" whose "eyes" were as unseeing as their "minds" were absent.
The point to be stressed here is that the object of Taylor's representation, the referent of the discourse, is not the Weimar Republic as such, but rather the "sham" which the Republic constituted. The metaphor of the "sham" is controlling in the sense that it provides the paradigmatic axis of the passage which sanctions the movement from the perception of external appearance to internal rottenness on the syntagmatic axis. The implicit structure of the relationship between outside and inside, appearance and reality, of the Republic is the same as that between the "eyes" and the vacant "minds" of its sympathetic "foreign observers." It is a form without a substance. And to characterize this substanceless form is the ultimate aim of the rest of Taylor's account of the history of Germany from 1815 to Hitler.
Now, I wish to suggest that a similar kind of analysis could be made of Taylor's whole book or, indeed, of any historical work, including those especially that we normally think of as "classics," such as works by Michelet, Ranke, Tocqueville, Marx, Burckhardt, as well as by such moderns as Hui-zinga, Braudel, Marc Bloch, and Croce. If anything, such works lend themselves to the kind of rhetorical analysis that I have attempted on this passage of Taylor's much more readily than his does, so manifestly "literary" are they. The passage from Taylor was chosen for what might be regarded as a kind of analytical overkill because it is so unself-consciously rhetorical, so patently intended to set forth the facts without embellishment, to present the argument crisply and in a straightforward way. My purpose has not been to cast doubt on the specific interpretation that Taylor offers of his materials,
[63]
but to explicate what might be meant by the "point of view" from which he wrote, and to show that what he says about his ostensible topic and how he says it are really indistinguishable.One could hardly praise the passage for the vividness of its language. Indeed, most of the metaphors contained in it are dead ones, but the appeal of dead metaphors to particular groups of readers should not be underestimated. They can, in fact, be comforting, having the effect of reinforcing views already held and serving to familiarize phenomena that otherwise would remain exotic or alien. It is seldom noted how the effect of "objectivity" can be attained by the use of non-poetic language, that is to say, by language in which dead metaphors rather than vivid ones provide the substance of the discourse. But whether dead or vivid, the language of this passage functions in precisely the same way that poetry does to move attention from the manifest level of the discourse to a latent or figurative level and back again. This earns the author the right to the formal explanation of why things are other than they seem to be and why his account reveals the way things really were.
Now, if by this point I have established the plausibility of the idea that every historical discourse has a figurative level of meaning, it is possible, I think, to resolve some conventional problems of historical theory. First of all, we can now see the similarities as well as the differences between "philosophy of history" and "historiography." Like any "philosophy of history," an historical narrative gains its effect as an explanation by its revelation of the deeper meaning of the events that it depicts through their characterization in figurative language. Their principal difference consists in the fact that whereas in "philosophy of history" the figurative element in the discourse is brought to the surface of the text, formalized by abstraction, and treated as the "theory" that guides both the investigation of the events and their representation, in the historical narrative the figurative element is displaced to the interior of the discourse where it takes shape vaguely in the consciousness of the reader and serves as the ground whereon "fact" and "explanation" can be combined in a relation of mutual adequation. Insofar, then, as the proper "historian" remains unaware of the extent to which his very language determines not only the manner, but also the matter and meaning of his discourse, he must be adjudged less critically self-conscious and even less "objective" than the "philosopher of history." The latter at least tries to control his discourse through the use of a technical terminology which makes his intended meaning clear and open to criticism.
Secondly, the disclosure of the presence of the figurative element in every historical discourse permits us to understand better the relation between a putatively "historical" mode of representation and that "historicist" mode which is its supposed antitype. Here we can observe that the crucial distinc-
[64] tion is not between an interest in the particular, as against the general, aspects of the historical process. It is rather between those writers of history who recognize that there is no choosing between these two aspects of the historical field, and those who think that such a choice is possible. The historical discourse seeks to explicate the relation between parts and wholes or between the phases and the completed structure of a process. In the absence of a specific theory of this relation, one is driven to utilize the tropes of language metaphor, metonymy, and synecdoche in order to figure it. These tropes may appear in a highly stylized, abstract form in "historicist" representations of reality, but they are no less present in "historical" representations as the "theory" guiding the articulation of the discourse. It does not matter whether the form of the discourse is that of the story well told or that of the logical treatise. In the absence of a genuinely scientific analysis of the modes of relationship obtaining among the elements of the historical field, tropology is the only conceptual protocol we have. In this respect, the presumed differences between a "narrative" and a "synchronic" account are more a matter of emphasis than of content. It may well be that the "historicist" characteristically uses particulars to exemplify or illustrate the general principles that he has claimed to discover in his study of history. But this does not mean that the "historian's" effort to concentrate on particulars "for their own sake" frees him from an appeal to those generalizations by which to weld his description of the particulars into a comprehensive narrative. This appeal is contained in the figurative language that the historian uses both to describe the elements in his field and to characterize the changes of this field during its process of development. The "generalizations" may be displaced to the interior of the discourse and the "particulars" placed in the foreground, but this secondary level of meaning serves the same function in the historian's discourse that theories do in that of the "historicist."Thirdly, this analysis of the figurative level of historical discourse permits us to conceptualize the possible types of historical representation by identifying the tropological mode which governs the figurative characterization of both the structure of a given historical domain and the phases of its articulation as a process. The tropes of metaphor, metonymy, synecdoche, and irony have long been recognized as constituting the principal generic types of figurative language-use. I have elsewhere tried to show19 that a given historian's style of representation can be characterized in terms either of his favoring one or another of these tropes, or of his efforts to mediate between them; my argument is that the form of the putatively straight historical narrative is as much dependent upon the governing tropological mode of figuration as the form of any historicist account is dependent upon the theory which it seeks
[65]
to justify. The various forms of historical narrative are as much a product of the effort to grasp the world figuratively as the various forms of historicist representation are determined by the theoretical apparatus of their different authors. My own intuition is that the various modes in which theory is articulated in the different sciences represent theoretical formalizations of the tropes of natural language. This is certainly true of those fields such as history which have not yet settled on a particular mode of language-use as the standard protocol for the description of data, the formulation of problems, and the reporting of their resolutions. Thus, I would argue, though I cannot defend the argument here, we can speak of the metaphorical, metonymic, synecdochic, and ironic modes of historical discourse. And because these modes correspond to the readers' modalities of language use (and therefore to their ways of conceptualizing the world), they provide the ground for the communication of understanding and meanings between specific "schools" of historians, on the one side, and specific publics, on the other. Because there is a generally poetic element in all historical writing, an element that appears in prose discourse as rhetoric, great historical works, whether by "historians" or "historicists," retain their vividness and authority long after they have ceased to count as contributions to "science."Fourth, recognition of the figurative dimension in the historical discourse provides us with a new perspective on the problem of historical relativism. The older, Rankean historicism was relativistic insofar as it believed that understanding of an historical phenomenon required that the historian view it "in its own terms" or "for itself alone." Here "objectivity" meant getting outside the historian's own epoch and culture, thinking his way into the consciousness of the age under study, viewing the world from its perspective, and reproducing the way the world appeared to the actors in the drama that he was recounting. The newer, absolutist branch of historicism that of Hegel, Marx, Spengler, et alia, those "scientistic" historicists castigated by Popper claimed to transcend relativism by the importation of scientific theories into historical analysis, use of a technical terminology, and disclosure of the laws that governed the historical process over all times and places. So too the more modern, social scientifically-oriented historians claimed to transcend relativism, by virtue of their use of a rigorous method and their avoidance of the "impressionistic" techniques of their more conventional narrativist counterparts. But if my hypothesis is correct, there can be no such thing as a non-relativistic representation of historical reality, inasmuch as every account of the past is mediated by the language-mode in which the historian casts his original description of the historical field prior to any analysis, explanation, or interpretation he may offer of it.
The use of a technical language or a specific method of analysis, such as, let us say, econometrics or psychoanalysis, does not free the historian from
[66] the linguistic determinism to which the conventional narrative historian remains enslaved. On the contrary, commitment to a specific methodology and the technical terminological system that it requires will close off as many perspectives on any given historical field as it opens up. It is not a matter, then, of choosing between the relativistic historicism of a Ranke and the more objective historicisms of Marx, Spengler, Weber, or Toynbee. Nor is it a matter of choosing between the new "social scientific" techniques of econometrics, psychoanalysis, or demography and the older, distended narrative techniques of the great storytellers of history. They are all equally relativistic, equally limited by the language chosen in which to delimit what it is possible to say about the subject under study.At the same time, however, if this theory of linguistic determinism is
correct, it offers a way out of an absolute relativism and a way of conceptualizing a
notion of progress in historical understanding. Because it is a theory of linguistic
determinism, we can envision a means of translating from one mode of discourse to another,
in the same way that we translate from one language to another. This way of
conceptualizing the problem of relativism is superior to that which grounds point of view
in epoch, place, or ideological allegiance, because we can imagine no way of translating
between these, while we can imagine ways of translating between different language codes.
It makes no sense to say that we can translate the perceptions of a Frenchman into
those of a German, those of a Renaissance man into those of a Medieval man, or
those of a radical into those of a liberal. But it does make sense to say that we can
translate the perceptions of an historian who has cast his discourse in the mode of
metaphor into those of one who has cast his in the mode of synecdoche, or those of one who
sees the world ironically into those of one who views it in the mode of metonymy. And if
the tropes of language are limited, if the types of figuration are finite, then it is
possible to imagine h