Recounting the Past: "Description," Explanation, and Narrative in Historiography


It is a rather widely held opinion among professional historians that the truly serious task of historiography, making it a contribution to knowledge and not a triviality, is the task of explanation. The opinion has roots in an objectivist myth that (as Peter Novick has recently shown) remains "powerful, and perhaps even dominant" within the historical profession.1 In the objectivist view, the "deнscription" of historical facts is unproblematic, as are the interpretive perspectives from which historians order these facts. Seconding the impact of professional historiography's founding myth is the continuing authority of a somewhat outdated, but still in many ways influential, philosophy of science. How many of us have downgraded a student paper with the words, "This is mere description"? The critical judgment may well be sound, but the choice of language is unfortunate. One rightly criticizes a work of history for being mistaken or uninteresting, but there is no warrant for assuming that "description" is a lesser historiographical aim than explanation. The assumption mistakes the character of historiography's contribution to knowledge; it rnisvalues narrative history in particular. The quality of a work is not adequately judged by its proportion of explanation to "mere description." Rather, a proper judgment of quality concerns all the aims that a work sets for itself. These are of four intertwined kinds: interpretive, "descriptive," explanatory, and argumentative or justificaнtory. It seems true by definition that every work of history embodies these aims. Different works embody them in differing degrees.

The argument I wish to make here is more theoretical than historical, even though I refer to empirical fact in articulating it. The aim is to counter


Earlier drafts of this essay were presented at the University of Iowa in a Faculty Rhetoric Seminar in July 1988 and in a Project on Rhetoric of Inquiry (POROlVhistory department'seminar in October 1988. I thank this Denkkotlektiv generally. Particular thanks for comments on the drafts go to Dudley Andrew, Frank Ankersmit, Bob Berkhofer, Steve Burton, Ken Cmiel, Jeffrey Cox, Evan Fales. Jan Goldstein, Jack Grant, Bob Hariman, Thomas Haskell, Steven Hoch, David Hollinger, Marcelyn Hutton, Hans Kellner, Ed Riser, David Klemm, Wallace Martin, Tom Mayer, Don McCloskey. John S. Nelson, Nancy Partner, Hunter R. Rawlings III, Paul A. Roth, Bill Schweiker, Barry Smith, Alan Spitzer, Serena Stier, Shel Stromquist, and Stephen Vlastos. I am indebted also to' my research assistants, Lori Brandt, Kevin Burnett, Ann Gallagher, and Sue Peabody.

1 Peter Novick, That Noble Dream: The "Objectivity Question" and the American Historical Profession (Cambridge, 1988)



professional orthodoxy. I investigate certain methodological contentions of Francois Furet. and Fernand Braudel, addressing their validity and playing them off against Braudel's historiographic practice in The Mediterranean and the Mediterranean World. Furet and Braudel are used only as examples, for the intent is to go beyond the practice of specific historians to address some conceptual underpinnings of historiography generally. We should strive to be constrained neither by the particular "craft" exemplars that we admire (whose very merit threatens to imprison us), nor by unexamined assumptions in our milieu, nor by restrictive philosophies of science, but only by the limits of inquiry itself. Awareness of conceptual underpinnings helps us to see beyond what was and is being done to what might be doneЧto what is allowed by a discipline beyond the particular practices of its past and present exponents. This is why theory is important.

The privileged place of explanation in our academic culture is most clearly manifested in the theoretical and methodological literatures. I use the term "explanation" not in the broad sense of "to elucidate" or "to make clear" but in the sense customary in philosophical and social science circles, where in most contexts "to explain" something means to say what caused it. To explain something in the terminology used here is to answer the question, "Why?" taking that question in the sense of, "What caused it?"2 Evidence of the privileged place of explanation, so defined, is to be found in the theoretical and methodological handbooks. Many announce explicitly their concern with explanation. Standard texts such as Arthur Stinchconibe's Constructing Social Theories are quite clear on this point.3 The handbooks are not always precise about what they mean by explanation, but usually the core if not the exclusive meaning is the answering of a causal question.4 Conversely, little has been written on "description," and in


2 I am of course aware of the somewhat problematic status of the notion of cause in the positivist tradition in philosophy, but this is not an issue that need concern us here. For reservations about "cause," see Bertrand Russell's classic article, "On the Notion of Cause" (1912Ч13), in his Mysticism and Logic, arid Other Essays (New York, 1918), 180-208. For a fastidious refusal to use directly the terms "cause" and "effect" combined with a constant invocation of these very terms, see Carl G. Hempel's equally classic "The Function of General Laws in History" (1942), in Patrick Gardiner, ed., Theories of History (New York, 1959), 344-56.

3 Among the works 1 have consulted are: Eugene J. Meehan, Explanation, in Social Science: A System Paradigm (Homewood, 111., 1968); Philippe Van Parijis, Evolutionary Explanation in the Social Sciences: An Emerging Paradigm (Totowa, N.J., 1981); Abraham Kaplan, Thi'.Conduct of Inquiry: Methodology for Behavioral Science (Scranton, Pa., 1964); Robert Borger and Frank Cioffi, eds., Explanation in the Behavioral Sciences (Cambridge, 1970); Patty Jo Watson, Steven A. LeBlanc, and Charles L. Redman, Explanation in Archaeology: An Explicitly Scientific Approach (New York, 1971); David Harvey, Explanaнtion in Geography (London, 1969); Paul Kiparsky, Explanation in Phonology (Dordrecht, 1982); Willem Doise, L'Explication en psychologic sociale (Paris, 1982); Peter D. McClelland, Causal Explanation and Model Building in History, Economics, and the New Economic History (Ithaca, N.Y., 1975); Christopher YXctyA, Explanation in Social History (Oxford, 1986). On the explanatory function of social theories, see Arthur L. Stinchcombe, Constructing Social Theories (New York, 1968), vii, 5, and passim.

4 One reason for this relative unanimity in preoccupation and definition is that the theoretical and methodological literature remains deeply influenced by the philosophy of science of the 1940s, 1950s, and 1960s. Philosophers of science in that generation were overwhelmingly concerned with "explanation," which they viewed as the answering of the "why" question. See Carl G. Hempel, Aspects of Scientific Explanation and Other Essays in the Philosophy of Science (New York, 1965), 245, 344,



many of the social and behavioral sciences "descriptive" types of research are viewed with suspicion.5 To be sure, in recent years there has emerged a newer theoretical literature, emphasizing the interpretation and description of culture, that turns the focus of attention back to other aspects of understanding than the explanatory.6 But the older, "harder" methodology still occupies the commandнing heights, the rhetorical high ground that the word "science" brings with it.7

To what extent, in their talking and thinking about historiography, do practicing historians share the view that explanation is the historian's central task? Eschewing a research survey of impossible subtlety, let us move instead by introspection, informed by several salient examples of the "bias for explanation."

In an essay in 1961, "Causation and the American Civil War," Lee Benson made use of E. M. Forster's distinction between "story" and "plot." A story is "a narrative of events arranged in their time-sequence," such as, "The king died and then the queen died." A plot is "also a narrative of events, the emphasis falling on causality," as in, "The king died and then the queen died of grief." Benson went on to gloss the distinction in this way:

Using Forster's criterion, we can define a historian as a plot-teller. Unlike the chronicler, the historian tries to solve the mystery of why human events occurred in a particular time-sequence. His ultimate goal is to uncover and illuminate the motives of human beings acting in particular situations, and, thus, help men to understand themselves. A historical account, therefore, necessarily takes this form: "Something happened and then something else happened because . . ." Put another way, the historian's job is to explain human behavior over time.8

Or consider E. H. Carr's assertion in What Is History? that "[t]he study of history is a study of causes" and his repeated characterization of a proper


and passim; Carl G. Hempel, Philosophy of Natural Science (Englewood Cliffs, N.J., 1966), 47, 49; Ernest Nagel, The Structure of Science: Problems in the Logic of Scientific Explanation (New York, 1961), 15-16; and Wolfgang Stegmuller, Problems und Resultate der Wissenschaftstheorie und Analytischen Philosophic, Band I: Wissenschaftliche Erkldrung und Begriindung (Berlin, 1969), 77.

5 As Miriam Schapiro Grosof and Hyman Sardy have pointed out, A Research Primer for the Social and Behavioral Sciences (Orlando, Fla., 1985), 114, referring in particular to the case study type of research. Such research, they suggested, has a preliminary status: it may suggest hypotheses for further research and may provide "anecdotal evidence to illustrate more generalized findings" (112). One exceptional work, written from a neopositivist perspective, that does deal with description is C. Behan McCullagh, Justifying Historical Descriptions (Cambridge, 1984).

6 See Clifford Geertz, The Interpretation of Cultures: Selected E.ssays (New York, 1973), esp. chap. 1, "Thick Description: Toward an Interpretive Theory of Culture," 3Ч30; Paul Rabinow and William M. Sullivan, eds., Interpretive Social Science: A Reader (Berkeley, Calif., 1979); Clifford Geertz, Local Knowledge: Further Essays in Interpretive Anthropology (New York, 1983); and Paul Rabinow and William M. Sullivan, eds., Interpretive Social Science: A Second Look (Berkeley, Calif., 1987).

7 For example, see the uneasy encounter of an anthropologist, committed to "description," with a physicist, committed to explanatory generalizations, reported by Renato Rosaldo, "Where Objectivity Lies: The Rhetoric of Anthropology," in John S. Nelson, Allan Megill, and Donald N. McCloskey, eds., The Rhetoric of the Human Sciences: Language and Argument in Scholarship and Public Affairs (Madison, Wis., 1987), 87.

8 Lee Benson, "Causation and the American Civil War," in Benson, Toward the Scientific Study of History: Selected Essays (Philadelphia, 1972), 81-82. See E. M. Forster, Aspects of the Novel (New York, 1927), 47, 130, and passim.



historical account as one that gives the reader "a coherent sequence of cause and effect."9

Consider finally David Hackett Fischer's contention in Historians' Fallacies that "history-writing is not story-telling but problem solving" and that historical narration is "a form of explanation."111

If the historian-reader of this paper finds that he or she is in essential agreement with these three statements, I have established that the reader shares a bias for explanation. The statements by Benson and Carr assume that the essential connections in a historical account are causal. The revealing of causal connections is what I (and they) define as explanation. Fischer's position is ambiguous, for his definition of explanation embraces elucidation generally, not just causal analysis. Nonetheless, his insistence that history is "not story-telling but problem solving" (a notion also advanced by Furet) seems to confirm the presence of an explanatory bias in my sense.''

Insofar as they are numerous, readers who find no disagreement with these statements confirm the claim by the historian and theorist of history Paul Veyne that "[t]here is ... a widespread idea that a historiography worthy of the name and truly scientific must pass from 'narrative' to 'explanatory' history." They also confirm the philosopher Paul Ricoeur's similar claim that in "history as a science . . . the explanatory form is made autonomous."12

The bias for explanation in much contemporary thinking about knowledge needs itself to be historically understood.

First, much of our thought about science is excessively influenced, even at this late date, by the history of Newtonian physics. Earlier in our century, the philosophical school variously known as "logical empiricism" or "logical positivism"13 took an apparent fact about that history and converted it into a principle. Physical science in the eighteenth and nineteenth centuries was marked by attempts to extend Newtonian theory to ever more phenomena. To show how yet another range of phenomena could be derived from Newton's laws was to give an explanation of them. The cutting edge of science, it seemed, was neither "descriptive" nor interpretive but explanatory. The payoff in physics did not come in the ordering of phenomena into descriptive types, as it did in natural history. Nor did it come in finding new ways of conceiving the physical


9 Edward Hallet Carr, What /.v History? (1961; 1 pt. edn., New York, 1967), chap. 4, "Causation in History," 113-43; quotes from 113 and 130; see also 111-12, 114, 138.

10 David Hackett Fischer, Historians' Fallacies: Toward a Logic of Historical Thought (New York, 1970), xii, 131.

11 In Historians' Fallacies, xv, n. 1, Fischer defined explanation as follows: "To explain is merely to make plain, clear, or understandable some problem about past events, so that resultant knowledge will be useful in dealing with future problems." Although Fischer did not notice it, there is a tension between the first and second clauses, since "usefulness] in dealing with future problems" suggests knowledge of cause-effect relations, hence a causal conception of explanation.

12 Paul Veyne, Writing History, trans. Mina Moore-Rinvolucri (orig. French edn., 1971; Middletown, Conn., 1984), 305 n. 5; Paul Ricoeur, Time and Narrative, trans. Kathleen McLaughlin and David Pellauer, 3 vols. (orig. French edn., 1983-85; Chicago, 1984-88), 1: 175.

1:1 On these terms, see David Oldroyd, The Arch of Knowledge: An Introductory Study of the History of the Philosophy and Methodology of Science (New York, 1986), 231, 248. Despite its slight inaccuracy, I shall use the more familiar term, "logical positivism," here.



world, for until the 1890s the Newtonian interpretive framework was thought to be beyond question.14

Second, within the context of the human sciences, a striking feature of secular, modernist: academic culture has been its commitment to metaphors of verticality, most evident in Karl Marx and Sigmund Freud but not limited to those traditions. It is a common trope of modernist inquiry that things more or less directly observable are not the "real" reality. In this view, the task of inquiry is to get down to what is hiddenЧto "underlying" determinants, to the "fundaнmental" features of the situation. The perspective of inquirer and audience has a crucial role in determining what will be regarded as insightful rather than as mistaken or simply irrelevant. Metaphors of verticality tend to grant a privilege to the explanatory project. David Hume's demonstration that we cannot observe causation underpins the view that explanation is tieferliegend (deeper) than description.15 When the philosopher of social science, Philippe Van Parijis, claimed that "any explanation assumes the operation of an underlying mechanism," he unwittingly reported the presence of this same metaphor."' Discussing Progressive social thought in America, Richard Hofstadter detected the assumpнtion that "reality" is "hidden, neglected, and off-stage," a similar trope with an identical function.17 When such metaphors are in place, the most striking insights will be those that claim to show how the "onstage" or "superstructural" things and events arise from previously invisible economic, sociological, or psychological conditions. These insights have an explanatory character, for they are answers to the question, "What caused it?"18

The cutting edge in a given discipline at a particular time may well be found in explanation. Base/superstructure metaphors are in no way contrary to the advance of knowledge, so long as they continue to produce new insights and so long as their heuristic, limited character is kept in view. But disciplines tend toward sclerotic self-satisfaction. Methodological rules articulated in one context


14 See Thomas L. Hankins, Science and the Enlightenment (Cambridge, 1985), 9, 20-21, 53; J. L. Heilbron, Electricity in the Seventeenth and Eighteenth Centuries (Berkeley, Calif., 1979), 6, 87 and following, 95 n. 47, 458 and following; and Christa Jungnickel and Russell McCormmach, Intellectual Mastery of Nature: Theoretical Physics from Ohm to Einstein, 2 vols. (Chicago, 1986), 1: xxiii and passim. On the classificatory impulse in natural history and elsewhere, see Hankins, Science and the Enlightenment, 113, 117, and passim; and Wolf Lepenies, Das Ende der Naturgescluchte: Wanclel hdtureller Selbstventandlichkeitcn in den Wissenschaften. des 18. uti/l 19. Jahrhunderts (Munich, 1976), 34. 47-48, 93, 98-102, 122-24.

15 Stegmiiller, Wissenschaftliche Erkliirung and Begriindung, 1. Kapitel, 2.b., "B'-rklarungen und Beschreibungen," 77.

16 Van Parijis, Evolutionary Explanation in Social Sciences, 6.

17 Richard Hofstadter, The Age of Reform from Bryan to F.D.R. (New York, 1969), 199-200. See also Lionel Trilling on V.L. Parrington and Theodore Dreiser in Trilling, "Reality in America," in his The Liberal Imagination: Essays on Literature and Society (New York, 1956), 3-21.

18 It seems plausible to suggest that thinkers less committed to the base/superstructure metaphor, or to other metaphors that envisage differentially visible realities, will be less committed to the explanatory project. One social science methodologist noted that most, of Max Weber's "theories" are actually "conceptual schemes and descriptions of'historical types'" (|ack P. Gibbs, Sociological Theory Construction [Hinsdale, 111., 1972], 16). It may well be that there'is a relation between Weber's well-known suspicion of the base/superstructure metaphor and the fact that his great achievements seem much more to be "descriptive" and interpretive than explanatory.



are often inappropriately applied to other contexts. Interpretive frameworks all too often come to be seen as die Sac he selbst.

Consider the bias for explanation as expressed in logical positivism. Of course, logical positivism has long been dead within philosophy. Killed by its own contradictions, it has given way to various neo- and post-positivisms. Nonetheнless, the formulations of logical positivism are important for two reasons: these express emphatically and with precision notions less clearly expressed elsewhere, and, second, many non-philosophers, including some historians, still cling to logical positivist dicta of three or four decades ago and trot them out whenever they want to appear rigorous and methodological.1''

In the first sentence of their widely cited paper, "Studies in the Logic of Explanation" (1948), Carl Hempel and Paul Oppenheim declared: "To explain the phenomena in the world of our experience, to answer the question 'why?' rather than only the question 'what?' is one of the foremost objectives of empirical science."20 In a similar vein, Ernest Nagel asserted in The Structure of Science: "It is the desire for explanations which are at once systematic and controllable by factual evidence that generates science; and it is the organization and classification of knowledge on the basis of explanatory principles that is the distinctive goal of the sciences."21 As a final example, consider the following assertion by logical positivist historians, appearing in a work that aspired to set the agenda for social science history in the United States, the Social Science-Research Council's "Bulletin 64": "The truly scientific function begins where the descriptive function stops. The scientific function involves not only identifying and describing temporal sequences; it also involves explaining them."22

None of these authors denies that "description" is part of empirical science; such a denial would, of course, be anti-empirical. But, by the same token, all consider "explanation"Чwhich they define essentially as I do hereЧto be "the truly scientific function." Given the rhetorical prestige that attaches to the term "science," we have no choice but to read these statements as manifestations of an explanatory bias.

Two mistaken prejudices supportedЧand continue to supportЧthe bias. One is the prejudice for universality; the other is hermeneutic naivete, or the belief in immaculate perception.

The prejudice for universality elevates explanation over "description" because in the logical positivist view "description" is tied to the merely particular, whereas explanation is seen as universalizable. In the immediate background to logical


19 The worst offender may well be economics: see Donald N. McCloskey, The Rhetoric of Economic-, (Madison, Wis., 1985), 7-8. But a survey in May 1987 by J. Morgan Kousser suggests that an "informal positivism," of an early Popper vintage, remains prevalent among historians: see J. Morgan Kousser, "The State of Social Science History in the Late 1980s," Historical Methods: A journal of Quantitative and Interdisciplinary History, 22 (1989): 12-20, at 14.

20 Carl G. Hempel and Paul Oppenheim, "Studies in the Logic of Explanation," in Joseph C. Pitt, ed., Theories of Explanation (Oxford, 1988), 9. On the importance of this article, see Ronald N. Giere, Explaining Science: A Cognitive Approach (Chicago, 1988), 28.

21 Ernest Nagel, The Structure of Science: Problems m the Logic of Scientific Explanation (New York, 1961), 4.

22 Social Science Research Council, Committee on Historiography, The Social Sciences in Historical Study [Bulletin 64] (New York, 1954), 86.



positivism stands the still remarkably influential opposition, first proposed by Wilhelm Windelband in 1894, between the "nomothetic" sciences, concerned with the search for general and invariable laws, and the "idiographic" historical disciplines, whose focus of attention is held to be particular entities.23 At least in principle, Windelband accorded equal status to nomothetic and idiographic investigations: in his eyes, both were science (Wis.senschaft). The logical positivists, in contrast, restricted the name and status of science to nomothetic investigaнtions, to those fields producing, or claiming to produce, general laws.

Because they often confuse "general laws" with other kitids of generalizations, historians sometimes miss the full force of the idea that a field is scientific only if it produces general laws. By "generalization," historians usually mean a broad statement that is nonetheless still tied to a particular historical context. In historians' language, the following invented statement counts as a generalization (the question of whether or not the statement is correct does not concern us here): "As a result of the growth of towns and trade, feudalism gave way to incipient capitalism in late medieval and early modern Europe." The "problem of generalization," as historians conceive of it, is usually the problem of how to get from fragmentary and confusing data to such larger assertions.24 But such assertions are not what the logical positivists, or Windelband before them, had in mind when they spoke of general laws. In "nomothetic" science, the desired generalizations transcend particular times and places, as in, for instance, this invented statement: "Whenever, within a feudal system, towns and trade begin to grow [we would likely find enumerated further conditions, along with stateнments concerning their interrelations], then feudalism gives way to capitalism." In short, the generalizations in question are laws (which can be formulated as "if. . . then" statements), and assemblages of such laws brought together in theories.

The Windelbandian distinction between the particular and the general has often been equated with the distinction between "description" and explanation. Consider the following passage, which begins Hempel's "Function of General Laws in History" (1942):

It is a rather widely held opinion that history, in contradistinction to the so-called physical sciences, is concerned with the description of particular events of the past rather than with the search for general laws which might govern those events. As a characterнization of the type of problem in which some historians are mainly interested, this view probably can not be denied; as a statement of the theoretical function of general laws in scientific historical research, it is certainly unacceptable.


23 Wilhelm Windelband, "History and Natural Science," with an introductory note by Guy Oakes, History and Theory, 19 (1980): 169-85, esp. 175. See also Georg G. Iggers, The German Conception of History: 'The National Tradition of Historical Thought from Herder to the Present, rev. edn. (1968; rpt. edn., Middletown, Conn., 1983), 147-52.

24 As in William O. Aydelotte's classic paper, "Notes on the Problem of Historical Generalization," in Social Science Research Council, Generalization in the Writing of History: A Report of the Committee on Historical Analysis, ed. Louis Gottschalk (Chicago, 1963), 145-77 (reprinted in Aydelotte, Quantification in History [Reading, Mass., 1971], 66-100).

25 Hempel, "Function of General Laws in History," 344Ч45.



As can be seen, at the same moment that he rejected Windelband's identification of historiography as idiographic, Hempel linked "description" to the particular. He did not argue for the linkage, but we can easily reconstruct why he believed that it was part of a view that "probably can not be denied." Consider again the two invented passages concerning the transition from feudalism to capitalism. The first has both a "what" and a "why" component. It is explanatory (or, more precisely, it claims to be explanatory), for it offers an account of why the transition from feudalism to capitalism took place. It is also "descriptive," for it says what was the case in late medieval and early modern Europe. But the second statement is different. It "describes" no reality. Rather, it states a set of hypotheses that are tied to no particular reality. Its relation is to concepts: feudal system, growth, cities, capitalism. When it is applied to a particular realityЧsay, Europe in the twelfth century or Lower Slobbovia in the twentiethЧit has an explanatory payoff, at least if the audience in question accepts the stated laws as true and agrees that the concepts in question are appropriate to that reality. "Why was there a transition from feudalism to capitalism in Lower Slobbovia in the twentieth century? Well, it is because whenever . . ." And so we have a form of explanation that has a portability, a universalizability, that "description" cannot have.

Given the prejudice for universality, the result is a general elevation of explanation over "description." It is widely held in philosophy and in social science that only knowledge of the general or universal (as distinguished from the local or particular) is truly scientific; all else is inferior. The prejudice has roots in Greek thought, in Plato and (even more influentially for science) in Aristotle. In his Metaphysics and elsewhere, Aristotle contended that knowledge of universals is the highest form of knowledge.26 In the Poetics, he noted the implication for history, observing that "poetry is something more philosophical and of graver import than history, since its statements are of the nature rather of universals, whereas those of history are singulars."27 In the twentieth century, the universalizing commitment is still alive, although in modern thought it derives more directly from Hume and from Immanuel Kant than from Aristotle. Poetry has dropped out of the circle of universal knowledge, which is now restricted to mathematics, natural science, and social science insofar as it follows the natural science model thus projected.28


26 Aristotle, Metaphysics, trans. W. D. Ross, 982'' 20Ч25, in Aristotle, Complete Works, ed. Jonathan Barnes, 2 vols. (Princeton, N.J., 1984), 2: 1554; on the "commitment to the generic" in Greek thought generally, see Windelband, "History and Natural Science," 181. There is, let it be noted, another side to Aristotle, exemplified in the Ethics and Rhetoric, where emphasis lies on specific cases (of moral judgment or of persuasion). (See Stephen Toulmin, "1 he Recovery of Practical Philosoнphy," American Scholar, 57 [1988]: 337-52, al 339 and passim.) But modernism looks with disfavor on the ethical-rhetorical strand.

27 Aristotle, Poetics, trans. Ingram Bywater, 14511' 6-7, in Aristotle, "Rhetoric" and "Poetics." introd. Edward P. J. Corbett (New York, 1984), 235.

28 On the idea of universality in modern thought, see, among others, Max Weber, "A Critique of Eduard Mever's Methodological Views," in Weber, The Methodology of the Social Sciences, trans. Edward A. Shils and Henry A. Finch (Glencoe, III., 1949), 163 n. 30; Stanley Rosen, Hermeneutics as Politics (New York, 1987), 45, 95 and passim; and Richard W. Miller, Fact and Method: Explanation, Confirmation and Reality in the Natural and the Social Scietices (Princeton, N.J., 1987), 3-4.



Recently, the prejudice for universality has been challenged on many fronts. The interpretivist strand has already been mentioned (note 6). The revival of rhetoric launches another, related challenge.29 Some methodologists of the social sciences have criticized excessive concern with universalizability.30 Even in philosophy of science, the idea that science is preeminently a matter of rinding "universal generalizations" has lately been questioned.31

Yet "interpretive social science" is still widely regarded as woolly headed, rhetoric is misdefmed as deceit, and challenges to the universalizability criterion in social science methodology and philosophy of science are not yet sufficiently appreciated. To be sure, few historians ever committed themselves to the positivist program of theory construction that, for instance, the Social Science Research Council urged on them in the early 1950s.32 In fact, the divergent practices of sociologically oriented historians and historically oriented socioloнgists have helped to reinforce the sense that the aims of historians are divergent from those of theory constructors.33 Even (perhaps especially) in social science history, the point now seems well understood.34 Nonetheless, logical positivism's "deductive-nomological" model of explanation (also known as the "covering law" model), which presupposes the pursuit of universal theory and seeks to explain


29 Among the works important for this revival are Chaim Perelman and Lucie Olbrechts-Tyteca, The New Rhetoric: A Treatise on Argumentation, trans. John Wilkinson and Purcell Weaver (orig. French edn., 1958; Notre Dame, Ind., 1969); Stephen Toulmin, The Uses of Argument (Cambridge, 1958); and Thomas S. Kuhn, The Structure of Scientific Revolutions, 2d edn., enl. (1962; rpt. edn., Chicago, 1970).

30 See Arthur L. Stinchcombe, Theoretical Methods in Social History (New York, 1978). esp. 1Ч4, 115-17.

31 See Giere, Explaining Science, 89-90, 102-04.

32 The authors of S.S.R.C. Bulletin 64 thought that "particular explanations of particular data . . . can contribute little to the cumulative growth of knowledge, unless used as the starting point for more systematic investigation and testing." Translated, the S.S.R.C. authors are saying that a study of, for instance, the Third Reich is not in itself a serious contribution to "the cumulative growth of knowledge." Only when such investigation yields a theory of, say, "fascist states" or "totalitarian political systems" has knowledge progressed. Consequently, the authors urged as a "general rule" that "problems should be defined and hypotheses developed as early in the analysis as possible," and they rejected "ad hoc hypotheses, drawn upon only after the evidence has been selected." The trouble with hypothesizing in the wake of, rather than before, serious examination of the evidence is that one's account begins to wiggle and squiggle in response to the facts and so comes to be a "mere description" of those facts instead of generalizable theory. See Social Sciences in Historical Study, 27.

33 The literature is large. But see, for example, the historian Olivier Zunz's comments on recent comparative-historical sociology, which he found "does not provide any real alternative" to the universalizing perspective of an earlier generation of historical sociologists, despite claims by its practitioners to have abandoned that perspective (Olivier Zunz, "Toward a Dialogue with Historical Sociology" [review of Theda Skocpol, ed., Vision and Method in Historical Sociology (Cambridge, 1984)], Social Science History, 11 [1987]: 31-41, quote at 38-39). In the Skocpol anthology, the historian Lynn Hunt acutely noted that the work of the historian and sociologist Charles Tilly has been at its best when it is either predominantly historical or predominantly sociological, rather than "caught uncomfortably between the two" (269). See also Victoria E. Bonnell's generally excellent "The Uses ot: Theory, Concepts and Comparison in Historical Sociology." Comparative Studies in Society and History, 22 (1980): 156-73.

34 Although there are exceptions. For example, one can detect, in Kousser's conviction that "rational choice theory" can serve as the basis for a synthesis in the field of political history that would in no way narrow the discipline's focus, a lingering attachment to older logical positivist hopes for synthesis through theory. (See J. Morgan Kousser, "Toward 'Total Political History,'" California Institute of Technology, Division of Humanities and Social Sciences, Social Science Working Paper, 581 [September 1985; revised November 1986], i, 17.) Let us inscribe on our walls the following observation: all calls for "synthesis" are attempts to impose an interpretation.



particular realities in terms of theory, remains influential in such fields as economics, sociology, and political science. Beyond these fields, the logical positivist program still retains an aura of prestige, partly because of the decisive way in which logical positivism seized control of the rhetorically powerful term "science" and partly because of its well-justified insistence on clarity and explicitness in inquiry.35

So the first mistaken reason for a general privileging of explanation over "description" is the prejudice for universality. The second, hermeneutic naivete, leads not to the elevation of explanation but to the debasement of "description." By hermeneutic naivete, I mean the viewing of the historical account as if it were a "view from nowhere," instead ofЧas it decidedly isЧa view from some particular interpretive perspective. Modernist academic culture, particularly when it claims the prestige of science, tends to repress the interpretive dimenнsion. Both Marx and Freud were notoriously prone to such repression, but their offense is far from unique. Once again, logical positivism provides an especially clear expression of a widely held view. Consider Hempel's "Function of General Laws in History." Many historian-readers will remember its centerpiece, the bursting of a car radiator. Hempel offered an explanation in a deductive-nomological form of the event, such that from certain initial and boundary conditions (for example, the bursting strength of the radiator metal, the temperature overnight) and from certain physical laws (for instance, concerning the freezing of water), the bursting of the car radiator can be deduced. The statement of initial and boundary conditions constitutes, of course, a "deнscription." Ironically, at the end of the essay, Hempel came to the proto-Kuhnian conclusion that the separation of "pure description" from "hypoнthetical generalization and theory-construction" is unwarranted.30 Presumably, then, every "description" is already permeated by "theory," as fact is by paradigm in Kuhn's image of science. Yet, in dealing with the radiator example, Hempel failed to take account of his own conclusion. Instead, he proceeded as if "pure description" were indeed possible.

Hermeneutic naivete is intertwined with the notion that "description" is intrinsically uninteresting. When the hermeneutic dimension is excluded, "deнscription" gets reduced to data collection. On this point, positivism holds to a position that most historians will recognize as faulty. Yet, even among historians of some sophistication, there remains a tendency to underrate the force and scope of the hermeneutic insight that all perception is perspectival. Richard J. Bernstein has usefully (if schematically) distinguished between pre- and post-Heideggerian notions of the "hermeneutic circle." In many standard characterнizations, the circle runs between part and whole within the reality that the investigator seeks to understand. For instance, a historian or a textual critic will come to understand one sentence in a document in light of the document as a whole. But in its wider, post-Heideggerian sense, the circle runs between the


35 See, on this last point, Lawrence Stone, "History and the Social Sciences in the Twentieth Century" (1976), in Stone, The Post and the Present (Boston, 1981), 16-18. 'Mi Hempel, "Function of General Laws in History." 356.



investigator and what is being investigated. The investigation will be prompted by the traditions, commitments, interests, and hopes of the investigator, which will affect what the investigator discovers. Conversely, the process of historical research and writing will change both the investigator and the audienceЧat least, it will do so if the inquiry is more than trivial.37 To come to grips with the interpretive aspect of inquiry, one must make a reflexive move, looking at the way that the inquirer's point of view enters into the investigation. The long historiographic tradition that holds to the fiction of an objective narrator feigning to be silent before the truth of the past resists self-reflexive sensitivity.38 The tradition goes along with an underrating of the "descriptive" project, which, as we shall see in relation to Braudel, is far more complex and interesting than a hermeneutically unaware perspective acknowledges.

It will not have escaped the reader's attention that I have enclosed the term "description" in quotation marks, the intellectual equivalent of rubber gloves. Unfortunately, the word is tied almost umbilically to the notion of "mere description"Чto the underrating of the project that it is meant to name. The project is the answering of the question, "What was the case?" as distinguished from the answering of the question, "Why was it the case?" (or "What caused it?") that is the hallmark of explanation. Given the infinite variety of perspectives from which a historical account can be written, both projects embody an infinite number of difficulties and possibilities.39

Accordingly, a term not so suggestive of the mere copying of some external model is called for. Thus I prefer the term "recounting" to designate answers to, "What was the case?" Linked to the French raconter, the term encourages us to think of this historical answering on the model of the telling of a taleЧin this case, a tale for the truth of which various arguments, documentary and otherwise, are made. There is clearly more than one way to tell a tale; by the same token, there are different ways of constructing the historical past. The new term helps us to appreciate that "description" is not a neutral preliminary to the real work of explanation, not mere data collection. It leaves us better able to see that the two cannot be given a differential importance in abstraction from the aims and audiences of particular historical works.

Those who miss the importance of recounting adopt (usually more or less unconsciously) one of two related positions. Either, while preserving a distinc-


37 Richard f. Bernstein, Beyofid Objectivism and Relativism: Science, Hermeneutics, and Praxis (Philaнdelphia, 1983), 135-36.

38 See Robert Finlay, "The Refashioning of Martin Guerre," and Natalie Zemon Davis, '"On the Lame,'" AHR, 93 (June 1988): 553-71 and 572-603, for an exchange that highlights the need to attend carefully to the different voices and attitudes that manifest themselves in a work of history. Even historians aware of the hermeneutic tradition often resist the self-reflexive implications. Note, for example, Quentin Skinner's apparently unwitting reduction of post- to pre-Heideggerian hermeneutics in "Hermeneutics and the Role of History," New Literary History, 7 (1975-76): 209-32.

39 As does the further argumentative or justificatory question: What grounds do we, author and audience, have for believing that such-and-such was the case, and that such-and-such is why it was the case? But I leave the justificatory question to the side in the present essay.



tion between recounting and explanation, they see recounting as uninteresting (as when it is taken as a mere preliminary to scientific knowledge), or they blend the two together but in such a way as to reduce recounting to explanation. In the two instances, the outcome is the same: an exclusion of recounting from the circle of valued knowledge.

The exclusion, moreover, is intimately bound up with questions concerning narrative and its validity. Narrative blends recounting and explanation. One of the effects of the bias for explanation and of the related bias for universality has been a debasement not just of "description" but of narrative. The celebrated "revival of narrative" has had to work against the still prevalent suspicion that "narrative history" is epistemologically and methodologically defective. When Lawrence Stone remarked that narrative "deals with the particular and specific rather than [with] the collective and statistical," it seems he was motivated in this assertion (which, as stated, is incorrect) by the uneasy thought that narrative is incapable of the theoretical universality that explanation by laws and theories promises, which would make it scientific. Thus narrative's alleged revival is shadowed by deeply held prejudices working against narrativeЧas they work against its primal component, recounting.4"

We can get at the questionable nature of these views by looking at their articulation by the Annales historian Furet, who dismissed both recounting and narrative for reasons closely connected to the philosophical and social scientific prejudices noted above. In "From Narrative History to Problem-oriented History" (1975), Furet discussed the advance of a new, analytical, conceptual, "problem-oriented" historiography and what he characterized as the "possibly definitive decline of narrative history."41 He approved of these developments, for narrative, he held, is logically and epistemologically flawed: "Narration's particular kind of logicЧpost hoc, ergo propter hocЧis no better suited to the new-type of history than the equally traditional method of generalizing from the singular."42 Admittedly, as a disabused positivist, Furet denied that the transition from narrative history to "problem-oriented history" suffices to bring history into "the scientific domain of the demonstrable." Such a goal, he suggested, is probably unattainable, but at least the transition brings history closer to it.4'

To what extent is Furet's characterization of narrative adequate to reality?


40 Lawrence Stone, "The Revival of Narrative: Reflections on a New Old History," Past and Present, 85 (November 1979): 3-24, quote at 3-4 (reprinted in Stone, Past and the Present, 74-96). Of course, even "the collective and statistical" does not rise to universality. "The collective and statistical" is itself "particular and specific." For what is being collected? Of what are the statistics gathered?

41 Francois Furet, "From Narrative History to Problem-oriented History," in Furet, In the Workshop of History, trans. Jonathan Mandelbaum (Chicago, 1982), 54-67: quote at 56.

42 Furet, "From Narrative History to Problem-oriented History," 57. See also In the Workshop of History, "Introduction," 8: "Traditional historical explanation obeys the logic of narrative. What comes first explains what follows."

43 On the unattainability of the goal, see Furet, "From Narrative History to Problem-oriented History," 66-67. On the connection of Furet's preference for problem-oriented history to the prejudice for universality, see In the Workshop of History, "Introduction," 6-7 (the new history becomes, as a "form of knowledge," applicable to any and all societies). See also "From Narrative History to Problem-oriented History," 60, where historical demography's transformation of "hisнtorical individuals" into "interchangeable and measurable units" also points to the presence of the universalizabilitv criterion.



Two points are of interest. First, like Stone, Furet alluded to narrative's supposed attachment to singulars, but, unlike Stone, he gave the attachment an explicitly negative cast by linking it to the empirical error of faulty generalizaнtion. The status of Furet's statement remains ambiguous, however, for he did not actually say (although his words appear to suggest) that narrative and generalizing from the singular have some special affinity for each other.

Much clearer is Furet's other assertion, that narrative follows the (il)logic of post hoc, ergo propter hoc. The same assertion has been made by some other writers as wellЧincluding the literary theorist Roland Barthes, whose own brief comнments on the allegedly fallacious character of narrative help to gloss Furet's rather clipped statement. In an influential essay, "Introduction to the Structural Analysis of Narratives" (1966), Barthes held that narrative is characterized by a "'telescoping' of logic and temporality": "Everything suggests . . . that the mainspring of narrative is precisely the confusion of consecution and conseнquence, what comes after being read in narrative as what is caused by: in which case narrative would be a systematic application of the logical fallacy denounced by Scholasticism in the formula post hoc, ergo propter hoc."44 Even though Barthes's statement may seem puzzling at first reading, the basic point is simple. Barthes is suggesting that narrative is a sequence of stated causes and effects. In short, he is making the same assertion abbut narrative that, above, we found Lee Benson and E. H. Carr making about historiography. By the same token, he is suggesting that narrative is essentially explanatory.

The causal/explanatory construal of narrative may be more familiar to some readers in the form given to it by the well-known American philosopher Morton White. In his Foundations of Historical Knowledge (1966), White contended that "a narrative consists primarily of singular explanatory statements," and that a history is "a logical conjunction of statements most of which are singular causal assertions." White distinguished history from chronicle, which is "a conjunction of noncausal singular statements." He then complicated matters by an explicit admission that a history may contain elements of chronicle and still be a history: this is why a historical narrative is only "primarily" causal or explanatory.45 But he did not go on to consider what impact the structure of chronicle might have on that of historical narrative. Implicitly, he thought of chronicle as "mere chronicle," just as historical "description" tends to bear the guise of "mere description." History proper is causal/explanatory.

The Barthes and Furet formulation of this idea is easily subjected to empirical test, for it makes a clear statement about the extant things that we call narratives. Narrative, Barthes suggested, confuses "consecution" and "consequence," leadнing us to see whatever it is that comes "after" X as being "caused by" X. This will indeed be the case if narrative is a chain of stated causes and effects, A causing


44 Roland Barthes, "Introduction to the Structural Analysis of Narratives," in his Image, Music, Text, trans. Stephen Heath (New York, 1977), 79-124, quote at 94. Barthes's statement is an intensification of Aristotle's assertion in the Poetics 1452'1 20 that "[t]here is a great difference between a thing happening propter hoc and post hoc."

45 Morton White, Foundations of Historical Knowledge (New York, 1965), 4, 14, 222-25, and passim; quotes from 4, 223, 222.



B causing C causing D, and so on. If narratives actually do invite their readers to equate consecutiveness with consequence, post hoc with propter hoc, it follows that narrative does function as a chain of causes and effects. Further, if this is the case, narrative will be adequately understandable in terms of the category of explanation alone. Conversely, if we do not find the post hoc, ergo propter hoc fallacy prominent in actual narratives, this will suggest the need for precisely that attention to nonexplanatory elements in narrative that the recounting/explanaнtion distinction encourages.

As it turns out, instances of causal-temporal confusion in narrative are fairly difficult to find. To be sure, in a perhaps unexpected narrative sphere, the cinema, Barthes's suggestion is illuminating, for it casts light on how viewers make sense of film action. When a camera shot shows one person pointing a gun, and the next shot shows another person falling to the ground and lying motionless, skilled viewers normally assume that the second person fell to the ground not only after the firing of the gun but also because of its firing. But the cinema is in some ways a special case of narrative, for there is usually no narrator's voice telling us the story; instead, the film feigns to show the story. In consequence, cinema seems to depend especially heavily for its coherence on viewer-inferred causal connections.40 In written fiction, it is difficult to find instances of causal-temporal confusion in the absence of a narrator of a certain sortЧone who, perhaps out of a stylistic commitment to parataxis, prefers to insinuate causal relations instead of stating them outright.47 Written fiction thus makes clear to us that causal-temporal confusion is not an essential part of fictional narrative but results instead from the narrator's adoption of a particular style of narration.

As for historiography, it can be clearly shown, contra Furet, that causal-temporal confusion arises not from the act of narration itself but from lapses in argument or justificationЧthe third aspect of the historical account, beyond recounting and explanation. Consider the following passage, from Nathan Rosenberg and L. E. Birdzell, Jr.'s Hoiv the West Grew Rich: "It is easy to imagine business enterprises formed among companions who learned to trust each other at war or at sea, for it happens often enough in our own times. (The generation which fought the American Civil War in their twenties, for example, invented the epitome of enterprises not based on kinship, the modern industrial corpo-


46 As Seymour Chatman observed, "it requires special effort for films to assert a property or relation" ("What Novels Can Do That Films Can't [and Vice Versa]," in W. J. T. Mitchell, ed., On Narrative [Chicago, 1981]: 117-36, at 124). On the showing/telling distinction, see Wayne C. Booth, The Rhetoric of Fiction, 2d edn. (Chicago, 1983), 3-20. On the application of Barthes's dictum to cinema analysis, see Michele Lagny, Marie-Claire Ropars, and Pierre Sorlin, Generique drs annees 30 (Paris, 1986), 25-26.

47 An example: "He had been chain-smoking for weeks. His gums bled at the slightest pressure from the tip of his tongue" (). D. Salinger, "For EsmeЧWith Love and Squalor," in Salinger, Nine Stories [New York, 1983], 104). Note how unstable the "confusion" is: the addition of an explicit "because" ("Because he had been chain-smoking for weeks") would be enough to destroy it. On the distinction between paratactic style, which does not spell out ranks and relations, and hypotactic style, which does, see Richard A. Lanham, Analyzing Prose (New York, 1983), 33-52.



ration, in their forties.)"48 In the second, parenthetical sentence, Rosenberg and Birdzell appear to be making two distinct statements. They tell us straightforнwardly that the invention of the modern industrial corporation followed the Civil War experience. At the same time, they insinuate that the invention of the modern industrial corporation was caused by the Civil War experience. Lay readers may well find nothing wrong with this piggybacking of an insinuation on an assertion. But competently trained professional historians,, when they enнcounter such a move, are likely to become suspicious and to ask for evidence. For example, how many of the inventors of the modern industrial corporation actually served in the Civil War? How close a connection can be drawn between such experience and their founding, two decades later, of corporations? What other factors might have prompted the development of corporations? The causal-temporal confusion in this text has nothing to do with the "particular logic" of narrative. It results from failure to adhere to a tacit rule in professional historiography against ambiguous assertion. One sees here an argumentative lapse, not the manifesting of an intrinsic property of narrative.

To sum up: narrative is not a scientifically disreputable application of the fallacy of post hoc, ergo propter hoc.49 This is not surprising. What is surprising is that a view contradicted by the reading of almost any good narrative historianЧ Thucydides, for exampleЧhas been asserted without serious challenge. Perhaps this indicates the depth of the bias for explanation. The sociologist Arthur Stinchcombe suggested that "[a]s the professional tone has taken over history (from the praising and damning tone . . . ), the normal linguistic effect is to make the narrative appear causal."30 Concluding that narrative is more than causal assertion, we are forced to attend to what is other than causal assertion in it. This means that we must attend to recounting.

Furet's attempt to deny to narrative history status as a legitimate form of knowledge production is closely connected to the distinction between narrative history and "problem-oriented history." But Furet did not originate the distincнtion: it was proposed by Fernand Braudel in the same year, 1949, that the first edition of his Mediterranean and the Mediterranean World appeared. Thus the stakes are larger than Furet: they concern a scientific mythology that has long


48 Nathan Rosenberg and L. E. Birdzell, Jr., How the West Grew Rich: The Economic Transformation of the Industrial World (New York, 1986), 125.

49 After completing my analysis of Furet's contention thai historical narrative follows the logic of post hoc, ergo propter hoc, I discovered that the philosopher W. H. Dray has also attacked Furet on this and other points: see W. H. Dray, "Narrative versus Analysis in History," in f. Margolis, M. Krausz, and R. M. Burian, eds., Rationality, Relativism and the Human Sciences (Dordrecht, 1986), 23-42, at 26 and following. Among other historians who, at one time or another, have linked narrative to a post hoc, ergo propter hoc logic are Lawrence Stone, Social Change and Revolution in England, 1540-1640 (London, 1965), xxii; and Charles Tilly, As Sociology Meets Hutoiy (New York, 1981), 90.

50 Stinchcombe, Theoretical Methods in Social History, 13. Of course, as will be evident from mv argument so far, tone is only part of the story. The explanatory bias derives more broadly from a certain view of science, from certain metaphors, from a concern in social science with pragmatic aims, and perhaps from other influences as well. Historiography does not exist in isolation from other intellectual and social practices.



shadowed (though fortunately never overpowered) the so-called Annales school. In a review of Charles-Andre Julien's book Les voyages de decouverte, Braudel articulated the contrast between an histoire-reat that "too often hides the background of economic, social, and cultural facts" and an histoire-probleme that "dives deeper [plonge plus loin] than events and men, a history grasped within the framework of a living problem or of a series of living problems clearly posed and to which everything that follows is subordinated, the joy of recounting [raconter] or of bringing the past back to life, the delights of making the great dead live again."51 How are we to characterize the histoire-probleme that Braudel recomнmended? The answer is offered by J. H. Hexter in his wittily parodic article on Braudel (1972). Histoire-probleme is history in which the asking of a "why" questionЧtaking that question in the sense of, "What caused it?"Чis uppermost in the historian's mind. In short, it is history that looks for explanations. As an example of histoire-probleme, Hexter cited Edmund Morgan's article, "The Labor Problem at Jamestown, 1607-18," which sets out to answer the question why, in a colony that by 1611 was on the verge of extinction, the colonists were to be found "at their daily and usuall workes, bowling in the streetes" instead of raising the crops needed to keep them alive.52

Hexter was forced to go to Morgan for an example because The Mediterranean and the Mediterranean World is not histoire-probleme.

First, there is no single, overriding causal question that the work poses. For example, it does not ask the question, "What caused 'the Mediterranean world' to come into existence?" Of course, even to think of the question is to recognize the extreme difficulty of answering it. What about causal questions of a more specific sort? Hexter cited three instances: "Why did banditry flourish in the Mediterranean toward the end of the sixteenth century?" "What accounts for the considerable flood of Christian renegades into the service of the Turk and the Barbary states?" "Why did the Spanish ultimately expel the Moriscos?"53 There-are many more, but seen in relation to the work as a wholeЧ1,375 pages in English translationЧthey play a relatively minor role. They appear intermitнtently. One will read for several pagesЧeven, exceptionally, for a dozen pages or moreЧwithout encountering an answering (or even an asking) of a "why" question. Then a question and perhaps an answer will appear. But one has no sense that the explanation, whether offered or only called for, in any way determines the general shape of the text. The explanations seem embedded in something much larger that is not explanation. For example, in the first three sections of Part One, Chapter 1, which take up sixty pages in the English text, I find only three clear instances of explanation-seeking questions.54 While Braudel


51 Fernand Braudel, "La double Failiite 'coloniale' de la Fiance aux XV et XVI1' siecles" (review of Charles-Andre Julien, Les voyages de decouverte et les premiers etablissements, XV et XVI'' siecles [Paris, 1948]), Annales: Economies, Soaetes, Civilisations, 4 (1949): 451-56, quotes at 452, 453.

52 J. H. Hexter, "Fernand Braudel and the Monde BraudellienЕ "Journal of Modern History, 44 (1972): 480-539, atа discussing Edmund S. Morgan, "The Labor Problem at Jamestown, 1607-18," AHR, 76 (June 1971): 595-611.

53 Hexter, "Fernand Braudel and the Monde Braudellien," 535.

54 Fernand Braudel, The Mediterranean and the Mediterranean World in the Age of Philip II, trans. Sian Reynolds, 2 vols. (New York, 1966), 1: 77, 82, 83.



posed such questions somewhat more frequently elsewhere in the book, the early sections are not greatly atypical.55

Second, it is a matter not simply of the intermittency with which Braudel offered explanations but also of the range of the explanations offered. The affinity between explanation and metaphors of verticality has already been noted. The metaphors are obviously present in Braudel's conceit that there exist three historical levels: the superficial, fast-moving, easily visible level of event; the more profound and slowly moving level of conjuncture; and the deepest geohistorical or structural level, which hardly moves at all and whose impact on human history is most easily missed.56 Moreover, he accepted the challenge that the conceit offers to the historian, of explaining by connecting one level to another. His most explicit statement of this aim occurs at the end of the preface to the second edition;57 it is also suggested in his review of Julien. And yet, as every serious commentator on The Mediterranean has observed, he failed to connect the different levels. Hexter noted that histoire-probleme provides an answer to the problem "of bonding event, conjuncture, and structure."58 But the answer is refusedЧto such a degree that the sociologist Claude Lefort, reviewing the work in 1952, saw in it a "fear of causality": "The condemnation of the causal relation leads [Braudel] into a pointillism that seems contrary to the sociological inspiration of the work."59

Braudel himself seems to have recognized that The Mediterranean did not fit the histoire-probleme mold. In the new introduction to Part Three written for the second edition, he suggested that recent research has made it possible for historians to choose from "two fairly well established 'chains'" in reconstructing the pastЧthe chain of economic events and conjunctures, and the chain of political events. A fully explanatory history would presumably reduce one chain to the other. But he went on to assert that "[f]or us, there will always be two chainsЧnot one."60 In the same introduction, he referred to the "bedrock of history" that is geography, and then immediately suggested that "the metaphor


55 A terminological note. Braudel frequently uses the word explanation in the broad sense of "to elucidate." He explicitly connects expliquer and eclairer in La Mediterranee el le monde raediterraneen a lepoque de Philippe II (Paris, 1949), 307 (the passage, in the introduction to Part Two, is omitted from the 2d edition). Accordingly, when Braudel uses "explanation," he does not always mean it in the sense in which the term is taken here. For one instance where it seems to mean "elucidation" without specific, reference to causes, see Mediterranean and the Mediterranean World, preface to the 1st edition, 20: "This book is divided into three parts, each of which is itself an essay in general explanation."

SB As Samuel Kinser pointed out, "Annaliste Paradigm? The Geohistorical Structuralism of Fernand Braudel," AHR, 86 (February 1981): 63-105, 83 and passim, Braudel changed his characterization of the first and second of these levels between the 1st and the 2d edition of Mediterranean and the Mediterranean World. Nor is Braudel necessarily committed to three levels in history. But these inconsistencies do not affect the basic point.

57 Braudel, Mediterranean and the Mediterranean World, 1:16.

58 Hexter, "Fernand Braudel and the Monde Braudellien," 535.

59 Claude Lefort, "Histoire et sociologie dans l'oeuvre de Fernand Braudel," Cahiers internationaux de sociologie, )3 (1952): 122-31; quote at 124. On the non-connection of levels in Braudel, see also Bernard Bailyn, "Braudel's GeohistoryЧA Reconsideration, "Journal of Economic History, 11 (1951): 277-82, at 279; and H. Stuart Hughes, The Obstructed Path: French Social Thought in" the Years of Desperation (New York, 1968), 58-59.

60 Braudel, Mediterranean and the Mediterranean World, 2: 902.



of the hourglass, eternally reversible" would be a "fitting image" of the work.61 In short, he himself deconstructed the metaphor of vertically that accompanies his notion of histoire-probleme.

To what genre, then, does The Mediterranean and the Mediterranean World belong if not to the professional genre of histoire-probleme? Following Braudel, Hexter suggested that it is "total" or "global" history.62 The characterization begs to be filled in. In another important contribution to the Braudel literature, Hans Kellner showed that the totalizing aspirations (inevitably unfulfilled) of The Mediterranean help to identify it as an "anatomy" or "Menippean satire." In his authoritative account of this literary form (the best-known manifestation of which is perhaps Robert Burton's Anatomy of Melancholy), Northrop Frye noted some of its most striking features: it engages in "dissection or analysis"; it is "loose-jointed"; it manifests "violent dislocations"; and it is apt, through the "piling up [of] an enormous mass of erudition," to turn into an "encyclopedic farrago" to which "a magpie instinct to collect facts is not unrelated."63 Even those who have only leafed through The Mediterranean should feel a sense of recognition. But the anatomy, as Frye also pointed out, is "a loose-jointed narrative form," manifesting "violent dislocations in the customary logic of narrative."Щ In short, The Mediterranean and the Mediterranean World is a work of narrative history.

It would be an understatement to say that The Mediterranean is not usually seen as narrative. But this is because "narrative" is usually taken to mean "the organization of material in a chronologically sequential order," to quote Lawrence Stone.65 Stone follows in a venerable tradition. His definition of narrative has roots in the Poetics, where Aristotle gives primacy to plot (muthos) over the other elements making up a tragedy.60 But if, as is usually done, we take "plot" to mean the sequence of actions within a work, the notion of plot focuses on only one aspect of narrative. "Action" implies an agent, and it also implies a setting within which action takes place. Accordingly, to make "chronologically sequential order" the defining feature of narrative is to engage in an arbitrary exclusion. To be sure, "traditional" historiography does tend to focus on action, and in consequence history has often been thought of as the story of actionsЧas the historia rerum gestarum. But we should not allow what is only an aspect of narrative to define narrative as a whole.

A century ago, Henry James called into question "the old-fashioned distinc-


61 Braudel, Mediterranean and the Mediterranean World, 2: 903.

62 On Braudel, Mediterranean and the Mediterranean World, 2: 1238, see Hexter, "Fernand Braudel and the Monde Braudellien" 530; see also 511.

63 Hans Kellner, "Disorderly Conduct: Braudel's Mediterranean Satire (A Review of Reviews)," History and Theory, 18 (1979): 197-222, reprinted in Kellner, Language and Historical Representation: Getting the Story Crooked (Madison, Wis., 1989), 153-87; Northrop Frye, Anatomy of Criticism: Four Essays (Princeton, N.J., 1957), 308-14.

64 FryCj Anatomy of Criticism, 309, 310.

65 Stone, "Revival of Narrative," 3.

66 Aristotle, Poetics, 1450a 2-17, in Aristotle, "Rhetoric" and "Poetics," 231. Even though, strictly speaking, the Poetics is about drama, not narrative, its influence far transcends such distinctions.



don between the novel of character and the novel of incident."'" The difference between the two extremes is a matter of degree, not of kind. We can imagine a continuum, running from fast-paced plots of incident ("Kojak" or "Miami Vice") to the novels of, say, Henry James. But the distinction between incident and character needs to be further broken down. Building on the Russian Formalist tradition, the theorist Seymour Chatman distinguished between action (carried out by an agent) and happening (an impingement on a character). One needs further to distinguish between character (which acts) and setting (which imнpinges). The interaction of the four elements produces the narrative. Two of the elements (action and happening) occur; two (character and setting) simply are. The first two we can call "events"; the last two, "existents." (Of course, existents can come into being, but this is no denial of the distinction between the emergence of an existent, which falls under the heading of event, and the existent itself.) Emphasis on one of the four elements perforce limits the attention given to the others. One might express this idea by means of a formula:

(AH) x (CS) = k

(action times happening [that is, "events"] times character times setting [that is, "existents"] equals a constant).1'8 It is simply tradition, when it is not uninformed prejudice, that insists on identifying narrative history with actions and happenнings, for characters and settings can also in principle serve as foci.

Accordingly, the crucial question to ask, in deciding whether a given work is best seen as an instance of narrative history, is not, "Is this text organized in a chronologically sequential order?" It is rather, "How prominent in the text are the elements of narrative?" In The Mediterranean and the Mediterranean World, they are prominent indeed, even though only Part Three, dealing with the "brilliant surface" constituted by political events, is chronologically ordered.m Succinctly put, The Mediterranean and the Mediterranean World is a work of narrative history that (except in Part Three) focuses not on events but on existents. Braudel turned the historical setting and the divisions and subdivisions of that setting into a vast collection of characters. These characters make up the single, all-embracing character that is "the Mediterranean and the Mediterranean world" itself.

Many of Braudel's commentators have pointed to his penchant for personiнfying. In an early review, Lucien Febvre remarked that Braudel promoted the Mediterranean to "the dignity of a historical personage." Hexter observed that Braudel populated the longue duree with "non-people personsЧgeographical entities, features of the terrain"; towns have intentions; the Mediterranean is a protagonist; even centuries are personalized. Kinser noted that Braudel treated


67 Henry James, "The Art of Fiction" (1884, 1888), in James, The Art of Criticism: Henry James on the Theory and the Practice of Fiction, eds. William Veeder and Susan M. Griffin (Chicago, 1986), 174.

68 See Seymour Chatman, Stoiy and Discourse: Narrative Structure in Fiction and Film (Ithaca, N.Y., 1978), 19, 32, 34, 44-45, 96-145. My formula is an expansion of one proposed by Wallace Martin, Recent Theories of Narrative (Ithaca, N.Y., 1986), 117-18, which in turn is inspired by a rather different formula in Gerard Genette, Narrative Discourse: An Essay in Method, trans, lane E Lewin (Ithaca, N.Y., 1972), 166.

69 Braudel, Mediterranean and the Mediterranean World, 2: 9(13.



space as "a human actor energetic and prompt to change costume."7" But we do not need to depend on the commentators, for Braudel himself was explicit about what he was doing. Consider the following passage, in the preface to the first edition: "Its character is complex, awkward, and unique. It cannot be contained within our measurements and classifications. No simple biography beginning with date of birth can be written of this sea; no simple narrative of how things happened would be appropriate to its history ... So it will be no easy task to discover exactly what the historical character of the Mediterranean has been."71 The Mediterranean and the Mediterranean World is best seen, then, as a vast character analysis, in which Braudel broke down "the Mediterranean," which begins as an undifferentiated entity, into its constituent parts, with growing attention over the course of the book to the human processes that are carried out within this geohistorical space. By the time he was through, "the Mediterranean" had become a massively differentiated entity. This is what we learn: that "ft]he Mediterranean speaks with many voices; it is a sum of individual histories," as Braudel wrote in the preface to the English edition (1972).72 The Mediterranean tells us what "the Mediterranean" was and, to some extent, what it still is. Braudel's explanations are contributions to this end. The work is a vast recounting, into which explanations are stuck like pins into a pin cushion. It is likewise a vast narrative, though more an anatomizing narrative of character than a sequential narrative of action.

The force and implications of this essay's distinction between recounting and explanation, and of its demonstration that Braudel's Mediterranean is in fact a work of narrative, are likely to be misunderstood by many readers. Some will carry in their minds awareness of recent polemics, heavily marked by political commitments, concerning the desirability or undesirability of "narrative history."73 Some readers will be inclined, wrongly, to see my attack on positivнism's a priori privileging of explanation over "description" as, in some way, a rejection of the legitimacy and importance of historians' explanatory efforts. Finally, some will misunderstand the nature of the distinctions that the essay poses. They are conceptual distinctions, aiming at clarity of thought about the historiographic enterprise. To say that a distinction can be made in thought is not to say that the elements thus distinguished will necessarily be clearly marked out in practice. In fact, the distinction between recounting and explanation is reader-constructed, but this is no denial of its reality, for the reader's active involvement with the text is a necessary condition of understanding.


70 Lucien Febvre, "Un livre qui grandit: La M editerranee el le monde mediterraneen a Vepoque de Philippe II," Revue hutonque, 203 (1950): 218; Hexter, "Fernand Braudel and the Monde Braudellien, 518-19; Kinser, "Annalist? Paradigm?" 67-68.

71 Braudel, Mediterranean and the Mediterranean World, 1: 17.

72 Braudel, Mediterranean and the Mediterranean World, 1: 13.

73 For a brief account of the controversy, with relevant references, see Novick, 'I'hat Noble Dream, 622-23; for a polemic against recent departures from "narrative history," see Gertrude Himmelfarb, The New History and the Old (Cambridge, Mass., 1987); and the AMI Forum on the old history and the new, which follows this article.



As suggested at the beginning of this essay, recounting and explanation are but two of the four main tasks of historiography.

Recounting some aspect of historical realityЧtelling what was the caseЧis the first task. A work in which this aim dominates will inevitably be ordered in narrative form, as denned hereЧthat is, historical actions, happenings, characнters, and settings will play, in varying proportions, a prominent role in the text.

Following on this is the explaining of some aspect of historical reality. If explanation becomes the historian's main concern, the work may well begin, in its focus on connecting explanans and explanandum, to depart from a primarily narrative form (although narrative does accommodate explanations).

Third, the historian claims that his or her recountings and explanations are true: otherwise, we would conventionally regard this scholar as something other than a historian. Thus the historical account has a third aspect, that of argument or justification. If the historian deals primarily with "the sources," the resulting account might well take the form of a commentary on, or analysis of, those texts. Alternatively, the purpose may be to justify a particular representation of the past against other possible representations, in which case the account would take on a primarily argumentative form. In both instances, the task of giving a narrative of the past would recede into the background.

Finally, the historian interprets the pastЧthat is, necessarily, views the past from some present perspective. The perspective permeates all that the historian writes: we have access to no regard de fin du monde, and even if we did it would be but one interpretation among others, God's interpretation as distinguished from all the rest. Since the historical account is necessarily written from a present perspective, it is always concerned with the meaning of historical reality for us, nowЧeven if, on an explicit level, it seeks to deny that it has any such concern.74 To the extent that the concern with present meaning is dominant, the historian becomes not simply a historian but a social or intellectual critic as well. Here, too, the historical account might well cease to be primarily a narrative of past existents and events.

The limits of the categorial schema need to be kept in mind. The claim is not that the categories enable a complete analysis of works of history but only that they chart out important dimensions of the historiographic enterprise.

In illustration of the recounting/explanation distinction, which is the focus of the present article, consider the following sequence of statements excerpted from Burns, Lerner, and Meacham's freshman college history textbook, Western CivilizationsЧa work that, both in the usual definition and in the definition offered here, is an instance of narrative history:

In 1839, along with the other great powers, Britain had signed a treaty guaranteeing
the neutrality of Belgium.

The Germans planned to attack France through Belgium.

[T]hey demanded of the Belgian government permission to send troops across its


71 For one such denial, see J. H. Hexter's charmingly naive essay, "The Historian and His Day," in Hexter, Reappraisals in HiiUny (Evanston, III., 1961), 1-13. The denial is of course central to the objectivist creed generally.



а(Belgium refused . . .

[T]he kaiser's legions began pouring across the frontier [anyway].

The British foreign secretary immediately went before Parliament and urged that his
country rally to the defense of international law and the protection of small nations.

[T]he [British] cabinet sent an ultimatum to Berlin demanding that Germany respect
Belgian neutrality, and that the Germans give a satisfactory reply by midnight.

The kaiser's ministers offered no answer save military necessity . . .

As the clock struck twelve, Great Britain and Germany were at war.75

Each of the nine statements tells what was the case. But, taken collectively, they are more than just a sequence of recountings, for they offer an answer to the explanation-seeking question, "Why did Britain go to war against Germany?" Once the reader has passed through the recountings, he or she will be positioned to see that the text offers an explanation as well. (One of the difficulties that weaker students have in reading such textbooks lies in their failure to make this leap.)

Explanation is dependent on recountings. To explain, as denned here, is to give an answer to the question, "What caused it?" In order to ask the question, we need an "it." Thus the question, "What was the case?" is primal: it precedes the explanation-seeking question. But the explanations offered will themselves be recountings. Assume that an audience has been brought to an elementary understanding of, say, the French Revolution. The audience has been offered an outline of the revolution: that it began in France in 1789 with the meeting of the Estates General, that its first important symbolic event was the Oath of the Tennis Court, that the Estates General became, soon thereafter, the National Assembly, that this was followed by the storming of the Bastille, that there was a war and a Terror, and so on. As part of this recounting, explanations of historical existents and events will be offered. The explanations, once accepted by an audience as persuasive, will become part of its image of what was the caseЧpart, that is, of what we might call a "representation" of the past. But images of what was the case always make possible further explanation-seeking questions. The further explanations, if accepted as persuasive, will enter into the image of what was the case and will make possible still more explanation-seeking questions.

Accordingly, what counts as an explanation in one rhetorical context may well count as a recounting in the next. The process is like the winning of land from the Zuider Zee. First, there is that part of the historical account that the audienceЧwhatever audience it is, amateurs or the most "advanced" profesнsional historiansЧsimply accepts as what was the case, not (or not any longer) calling it into question. This is like land won from the Zuider Zee and now solidly under cultivation. Second, there is that part that the audience is inclined to ask further explanation-seeking questions about. This is like the present shoreline of the Zuider Zee. Persuasive answers to explanation-seeking questions are like


75 Edward McNaJl Burns, Robert E. Lerner, and Standish Meacham, Western Civilizations: Their History and Their Culture, 10th edn. (New York, 1984), 927-28.



pumps and dikes that will turn this part, too, into dry landЧinto what is accepted as what was the case. There is next that partЧnot knowledge but nescienceЧthat is too far from accepted recountings to permit explanation-seeking questions but which may become an object of explanation in the future. Here we have the center of the Zuider Zee, hidden beneath the waters. Finally, not to be forgotten, there is the wider society within which historians write. This is like the North Sea, whose storms may invade the dikes and inundate part or all of what had been won, with apparent security, for cultivation. When this happens, the old recountings, and the explanations subsidiary to them, will come to seem mistaken, or at least irrelevant to important concerns of the present. In response to the recession of their hitherto implicit persuasiveness, a revision of the past will come to be demanded.

Yet, for all the interweaving of recounting and explanation, the distinction between them is justified and important. Consider another passage from Western Civilizations:

The Coming of the Revolution

Faced with serious challenges to centralized power from the resurgent noble elites as well as popularly based political movements in the eighteenth century, only the ablest absolutist ruler, possessing in equal measure the talents of administrative ability and personal determination and vision, could hope to rule successfully. The French king, Louis XVI, possessed neither of these talents. Louis came to the throne in 1774 at the age of twenty. He was a well-intentioned but dull-witted and ineffectual monarch . . .

Conditions in France would have taxed the abilities of even the most talented king; for one with Louis XVI's personal shortcomings, the task was virtually insurmountable. Three factors, in particular, contributed to the breakdown that produced revolution.76

Clearly, on one level, the passage offers us a recountingЧa statement of what the authors believe was the case in France in the period preceding the French Revolution. But, on another level, they are beginning to give an explanation why the revolution occurred. While the distinction between recounting and explaнnation is not always clearly marked within the historical text, a clear marker is present here in the form of a "contrary-to-fact conditional," or "counterfactual." As philosophers have long known, a statement about causation implies a counterfactual statement. When a historian states that C caused (led to, occaнsioned, brought about) E, he or she is simultaneously implying that without C there would have been no E, all other things being equal.77 In telling us that "only the ablest absolutist ruler . . . could hope to rule successfully," the authors explicitly introduced the counterfactuality that is present at least implicitly in all


76 Burns, Lerner, and Meacham, Western Civilizations, 674.

77 There is a large literature. For a convenient introduction to philosophers' discussion of causation and counterfactuals, see Myles Brand, "Causality," in Peter D. Asquith and Henry E. Kyburg, eds., Current Research in Philosophy of Science: Proceedings of the P.S.A. Critical Research Problems Conference (East Lansing, Mich., 1979), 252-81, at 264-69; see also the index entries for counter-factual conditionals in, among other studies, J. L. Mackie, The Cement of the Universe: A Studi of Causation (Oxford, 1980); and Wesley C. Salmon, Scientific Explanation and the Causal Structure of the World (Princeton, N.J., 1984). Of more direct interest to historians is the chapter "Counterfactuals and the New Economic History." in Jon Elster, Logic and Society: Contradictions and Possible Worlds (Chichester, Eng., 1978), 175-221.



explanation. Historians who remain unaware of how explanation, in its appeal to contrary-to-fact conditionals, differs from recounting tread on shaky epistemo-logical ground.

Recounting and explanation do not subsist alone; rather, they fit within the fourfold matrix suggested above. Often in historiological discussion, a distincнtion is made between "narrative" and "analytic" history. But the narrative/ analysis dichotomy is too crude to contribute much to understanding. Braudel's Mediterranean shows that some narrative is heavily analyticЧthat is, it engages in the differentiation of hitherto undifferentiated entities. Conversely, much analysis proceeds in (conventionally) narrative form, following "chronologically sequential order": a model instance is Marx's Class Struggles in France.7S The term "narrative" is used confusedly in contemporary theoretical discussion, although the confusions cannot be unpacked here. As for analysis, it takes place in quite different intellectual contexts, established by the four tasks of recounting, explanation, justification, and interpretation.

We have seen already that analysis can occur in the context of recounting. It occurs also in the context of explanation: thus Marx's detailed analysis of the class structure of French society in 1848 aims at explaining why the French revolution of 1848 turned out as it did. Finally, it occurs in the contexts of justification and interpretation, although when these dimensions are dominant the writer-inquirer is likely to be viewed as a textual critic or social critic respectively, rather than as a historian.

Related to the narrative/analysis distinction is the distinction between "narнrative" and "problem-oriented" history that Furet developed out of Braudel. In "From Narrative History to Problem-oriented History," Furet seems to imagine a breaking free of the latter from the former. In the introduction to In the Workshop of History, he complained that the British historian of France, Richard Cobb, "turns history into a laboratory for a purely existential preference." Hating "ideas" and "intellectualism," Cobb transforms the quest for knowledge "into a passion for novelistic narrative." Lacking "intellectual constructs," he is a social historian for whom "only individuals exist." His narrative is guided by a sympathy for the "life" of the period he describes. Sympathy, which replaces "the explicitly formulated question" as a guide to research, "belongs to the realm of affection, of ideology, or of the two combined." Thus history a la Cobb "remains purely emotional," failing to maintain "cultural distance between the observer and the observed." The product of such a history is "erudition"Чnot, we are given to understand, the true seriousness of a "problem-oriented history that builds its data explicitly on the basis of conceptually developed questions."79

Yet, in his neopositivist commitment to a universalizable (or at least a comparable) history that will supersede the current "proliferation of histories,"80


78 Karl Marx, The Class Struggles in France: 1848 to 1850, in Karl Marx, Surveys from Exile, ed. David Fernbach [Marx, Political Writings, vol. 2] (New York, 1974), 35-142.

79 Furet, In the Workshop of History, "Introduction," 13-20. The final quotation is from a version of Furet's introduction that appeared as "Beyond the Annales," Journal of Modern History, 55 (1983): 389-410, quote at 409; in the book, Furet recommended "an intellectualist history that builds" (20).

80 Furet, In the Workshop of History, 16.



Furet swept under the rug the fact that, on nicely "conceptual" grounds, explanation cannot be autonomous. Moreover, like many in the positivist tradition, he has forgotten that the explanatory theories that he wants historians to deploy presuppose particular interpretive standpoints that the theories themselves do not bring to light. Recountings (and explanations as well) must be carried out from some place, for some motive. The interpretive dimension is thus inescapable. The first words of The Mediterranean and the Mediterranean World are telling on this point: "I have loved the Mediterranean with passion."81 Braudel's words are as "affective" as anything in Cobb and his history as "erudite" as anything Cobb has written. These facts might be taken as excluding The Mediterranean from the true ranks of disciplinary history. In a review in 1953, Bernard Bailyn criticized the book for being "an exhausting treadmill," ruined by the fact that "[t]here was no central problem Braudel wished to examine," painfully lacking in "proper historical questions."82 But precisely at issue is what constitutes a "proper historical question."83 To focus on explanation alone is to exclude this issue: and yet it perpetually returns.

To say that explanation presupposes recounting is to say that it presupposes a presentation of narrative elements. But historiography is a collective enterprise, and it is quite possible for an individual historian to forgo, in greater or lesser degree, the telling of a narrative that is already largely known. Indeed, such a procedure often seems necessary if historical knowledge is to advance. To the extent that a basic narrative is not told but presupposed, the elements of narrative will tend to fade into the background. In such cases, there is a genuine departure from narrative history. Thus, in rejecting the narrative/analysis contrast, I am not making the essentially empty gesture of declaring all history to be narrative history.

The historian who is clearest on this matter is Alexis de Tocqueville. Consider the beginning of The Old Regime and the French Revolution: "The book that lies before you is not at all a history of the Revolution, for that history has been written with such success that I cannot dream of doing it again; instead, it is a study on the Revolution."84 Tocqueville is true to his word. Time and again in The Old Regime and the French Revolution, he referred to historical events and existents without recounting them in detail, relying instead on the reader's knowledge of them. His relative neglect of recounting freed him to move


81 Braudel, Mediterranean and the Mediterranean World, 1: 17.

82 Bailyn, "Braudel's GeohistoryЧA Reconsideration," 279, 281.

83 Compare Bernard Bailyn, "The Challenge of Modern Historiography," AHR, 87 (February 1982): 1-24, at 5: "Braudel's Mediterranee . . . should be known . . . for its ahistorical structure, which drains the life out of history. For the essence and drama of history lie precisely in the active and continuous relationship between the underlying conditions [my italics] that set the boundaries of human existence and the everyday problems with which people consciously struggle." How does Bailyn know that this is where the "essence and drama" of history lie?

81 Alexis de Tocqueville, L'ancien Regime et la revolution, in Tocqueville, Oeuvres, papiers et correspondances, ed. J.-P. Mayer (Paris, 1951- ), 2, 1: 69: "Le livre que je publie en ce moment n'est point une histoire de la Revolution, histoire qui a гte fake avec trop d'eclat pour que je songe a la refaire; c'est une etude sur cette Revolution." In the standard English-language edition, The Old Regime and tlw French Revolution, trans. Stuart Gilbert (Garden City', N.Y., 1955), vii, Tocqueville's point is slightly obscured.



forward on the three remaining fronts. He addressed head-on the explanation-seeking question, "What caused the revolution?"83 He argued explicitly against that representation of the revolution that saw it as essentially an attack on religious and political authority. He was likewise explicit about the interpretive dimension of the book and hence about the social criticism that it offers. As he observed, "I have never quite lost sight of present-day France." Thus, among other things, he sought to throw into relief "those virtues so vital to a nation" that he found absent from the contemporary scene but present earlier.86

Much of the hostility, often quite visceral, that in recent years has been directed against departures from "narrative history" seems actually to reflect an uneasiness about the knowledge-expanding capacity of an academic discipline. There is in some quarters a longing for the repetition of old pieties, the careful burnishing of myth. Physics, psychology, and even sociology seem largely protected from such expectations, but historiography, it appears, is not. It may seem tempting, in the face of such challenges, to take refuge in a sophisticated neopositivism that would stress the specifically explanatory task of historiograнphy over its other tasksЧfor we are inclined to see explanation as somehow insulated from issues of value.

The bias for explanation extends well beyond those historians influenced by positivist theory and methodology. For example, in a widely discussed essay, a neo-Hegelian intellectual historian told us that intellectual history "must address the issue of explanation, of why certain meanings arise, persist, and collapse at particular times and in specific sociocultural situations."87 Of course, this particular exercise in explanation, and explanation generally, is certainly part of what intellectual historians do. But to privilege the explanatory dimension is to put into the background the framework of assumptions that every explanatory project presupposes. These assumptions derive from the historian's own tradiнtions, commitments, interests, and experience, which finally cannot be histori-cized, cannot be subordinated to an objective, authoritative representation of history. The conservative critics of historiography are correct: history is about values. Historians qua historians, given the largely unreflexive character of their discipline, do not seem especially well equipped to deal with this fact.

Nonetheless, historians can at least know what they are doing when they are contributing to knowledge. It is not simply that they explain, as some contend. On the contrary, they first of all recount, in delight or fascination or horror or


85 See especially the string of "why" questions that he poses in his foreword, and his final chapнter, "How, Given the Facts Set Forth in the Preceding Chapters, the Revolution Was a Foregone Conclusion" (Tocqueville, Old Regime and the French Revolution, x, 203-11).

m Tocqueville, Old Regime and the French Revolution, xii. I write in the margins of Furet's highly intelligent and illuminating analysis of Tocqueville's book: Francois Furet, "De Tocqueville and the Problem of the French Revolution," in Furet, Interpreting the French Revolution, trans. Elborg Forster (Cambridge, 1981), 132-63. Characteristically, Furet tended to conflate the interpretive dimension of Tocqueville's project, concerned with "the meaning of his own time," with the task of articulating "explanatory theory" (132-33; see also 159-60). But, as I have argued, these are two distinct (though related) projects.

87 John E. Toews, "Intellectual History after the Linguistic Turn: The Autonomy of Meaning and the Irreducibility of Experience," AHR, 92 (October 1987): 879-907; quote at 882. Compare David Harlan, "Intellectual History and the Return of Literature," AHR, 94 (June 1989): 581-609.



resignation. Upon recountings, explanations arise.K8 Recountings and explanaнtions presuppose an interpretive perspective, and, in the best histories, they modify and enrich such a perspective. The articulation of perspectives is a contribution to knowledge that historians too often overlook or view with discomfort. To these tasks, justification stands as a sine qua non. Thus we see the legitimacy of narrative and, more generally, the legitimacy of forms of knowlнedge often deprecated in social science. The argument is not for the maxim "anything goes." It is rather for a critical pluralism, for standards of evaluation appropriate to the forms of knowledge being sought. It is a contribution to rigor, not a detraction from it.

88 See, again, Herodotus, The History, trans. David Grene (Chicago, 1987); and Thucydides, History of the Peloponnesian War, trans. Rex Warner (Harmondsworth, Eng., 1954). Herodotus was more inclined to become caught up in the sheer fascination of what he tells, while Thucydides leaned more toward explaining things. But both historians did both.


The American Historical Review, Vol. 94, No. 3 (Jun., 1989), pp. 627-653

Hosted by uCoz