New Literary History 27.3 (1996) 529-543

Literary History and the Search for Certainty

Robert Ellrodt


Is certainty attainable in science, in history, in meaning, and in literary theory? In the present intellectual climate many epistemologists, philosophers, historians, and literary critics are likely to give a negative answer. Some of them seek support for their skepticism in developments which occurred in the science of nature, quantum physics, and the theory of relativity. More recently the scientific use of such notions as randomness, turbulence, and chaos has been assumed to bring evidence of a total lack of order in the universe and justify the deconstructionist claim of indeterminacy in writing. 1 The picture, however, is different when we listen to what the scientists have to say. The following remarks are borrowed from a mathematician, André Lichnerowicz, and two physicists, David Ruelle and Bernard d'Espagnat, who cannot be suspected of clinging to positivist and intransigently deterministic views. 2

A distinction between different levels of reality is the first requirement. Newtonian mechanics still govern our everyday lives and our techniques though the notion of curvature has superseded the law of gravitation in interstellar space. Only in the microscopic infinite of quantum physics are quantities, when measured, affected by the instrument used. The uncertainty principle of Heisenberg shows we cannot determine simultaneously both the position and the velocity of an electron, but it will not deflect a bullet. Quantum mechanics even allows prediction in its own sphere. Scientists calculate the radiations of molecules and atoms accurately, and their evaluations, though based on amplitudes of probability, "are verified by experience every day and every where" (Rr 75). "If it pleases you to say that quantum mechanics is deterministic," says Ruelle, "it is. If you prefer to describe it as probabilist, you may say so, for it only asserts probabilities. . . . But it is not probabilist in the usual sense." Even in the aleatory oscillations of chaotic phenomena "the disorder of chance is created by a deterministic order" (Hc 128).

To make accurate predictions about some natural phenomena is still--and may remain--impossible if the beating of the wings of a [End Page 529] butterfly, as we are told, may perturb the weather forecast. Such limitations do not affect the certainties we have. They do not amount to total certainty, and probably never will. In a different sphere of man's quest John Milton in Areopagitica reminded us that the friends of Truth, imitating the careful search that Isis made for the mangled body of Osiris, "went up and down gathering limb by limb still as they could find them." "We have not found them all," he added, "nor shall ever doe, till her Masters second comming; he shall bring together every joynt and member, and shall mould them into an immortall feature of loveliness and perfection." 3

What modern science does not pretend to do is to disclose ultimate reality: like phenomenology it stops short at appearances. It describes our relations with nature, not nature in itself. Yet many physicists, d'Espagnat asserts (Rr 60), think that the existence of a structured reality behind appearances is the hypothesis that can best account for the regularity of our observations. And David Ruelle agrees that "there is in the laws of nature a greater harmony than one would have dared to expect" (Hc 170).

The method of science is described as hypothetico-deductive. From a hypothesis, consequences are deduced and the scientist devises experiments to produce them. If the expected results are observed, the hypothesis may be said to work, which does not mean that it is definitively confirmed, for other hypotheses might have led to the same consequences. That is why theories change, but a theory, once satisfactorily tested, retains part of its validity. The theory of Einstein displaced that of Newton, but Heisenberg himself declared the Newtonian law of gravitation still valid for all speeds inferior to the speed of light (Vp 247). Science is cumulative.

This is what socio-epistemologists vainly seek to deny. They cannot deny either that the rules for the prediction of observations are universal, not particular to one type of society. When Jean-François Lyotard reduces science to a "consensus between experts" or to a "legitimation by power" since the development of the techniques necessary for accurate observation depends on the support of the state or the community, 4 he leaves out of consideration the essential source of legitimacy: the verification of the hypothesis. When he assumes that to be "formally satisfying" for the experts is the only criterion for truth and gives as an instance the mathematics of the Bourbaki group, he chooses an example related to logic: Bourbaki mathematics are not used to send rockets to the moon. Mathematics at large are perhaps more radically nonontological than physics, yet, as Lichnerowicz points out, though mathematical models do not "explain" the real, they allow us to make predictions and control phenomena. Truth can no longer be described [End Page 530] as adaequatio rei et intellectus, but the difference between truth and error is not abolished. No scientist would accept the blunt statements of Greimas: "True and certain knowledge is only a matter of trust," "Belief and knowledge are dependent on one and the same cognitive universe." 5

I therefore intend to examine the possibility and limitations of certain knowledge in history, in the determination of meaning and in literary theory.

From the beginning different emphases have been placed on history in relation to truth and rhetoric. It was commonly treated as a branch of rhetoric in antiquity and in the Renaissance 6 and, as late as the seventeenth century, Descartes and La Mothe le Vayer complained "Du peu de certitude qu'il y a dans l'histoire," anticipating the modern trend which links history and fiction. Yet Aristotle had not placed history in his Rhetoric and from Herodotus to Bodin it was defined as narratio vera. To tell the truth was the first duty of the historian even in the eyes of Cicero and Quintilian. 7 Gassendi, in opposition to Descartes, developed an empirical view of historical truth based on the reliability of information, the logical correctness of interpretation and the balance of the soul required in the historian not to be disturbed by the affections. History so conceived could make use of rhetoric as the art of the probable in order to link men together in a common awareness of their past. 8

I have alluded to these conflicting views because they forestalled the modern debate and Gassendi's position is still a sensible middle way between extremes. One extreme is best represented by Hayden White's assertion that historical narratives are mere "verbal fictions the contents of which are as much invented as found, and the forms of which have more in common with their counterparts in literature than they have with those in the sciences." 9 Roland Barthes and Michel de Certeau claimed that the historian does not gather facts but "signifiers." 10 Hans Robert Jauss more moderately maintained that res factae and res fictae are not separable. 11 Why slur over the wide difference between the invented facts of fiction and the recorded facts of the historian? In some cases authenticated facts are not available but the very possibility of a discrimination between verified, doubtful, and unverifiable facts allows a scientific approach. Momigliano rightly spoke of "the invention of history" as only "a modish cliché" (L'he 40-41). 12

The facts, some say, are only "traces." The historian is never in front of the object he is supposed to study. But Ricoeur rightly maintains that the apprehension of the past in its documentary traces is an observation in the scientific sense of the term. 13 Historical facts, of course, must be interpreted and Barthes pointed out that they function like "indices." But so do the facts in a medical diagnosis. The true dividing line [End Page 531] between history and the physical sciences is that experimentation is impossible. Does it imply that the historian cannot establish any law?

One may agree that the so-called "covering-law model" proposed by Carl G. Hempel is bound to remain an unfulfilled ambition. Can any his-torian prove that whenever an occurrence is verified at a certain place and time, an occurrence of another specific type will be produced at a time and place related to the first event? 14 Such attempts, as Ricoeur observes, have only led to generalities which are pseudo-laws, for even their probability cannot be tested (with a possible, yet still doubtful, exception, for demography and economics). 15 One can at best expect high-frequency rather than invariable relations. Yet the historian is not thereby condemned to "the impressionistic approach" accepted by Philippe Ariès and paraded by Paul Veyne. 16

The historian may first aim at establishing particular causal connections not necessarily dependent on a law. Like the physician he has to rely on conjectural reasoning to identify the necessary and sufficient cause of the phenomenon. In a case involving human decisions Ricoeur admits that he can only claim the kind of probability expected from tragedy by Aristotle: "what a man would say or do necessarily or in all likelihood" (Tr I.185). The historian, however, may go further when he constitutes series of phenomena: political, economic, cultural. And by taking geographic conditions into consideration, he may be able to determine dominant traits over long periods, a method best illustrated by the works of Braudel. 17 LeRoy-Ladurie has cultivated in a different way what he called "l'histoire immobile" in his inaugural lecture at the Collège de France.

To bring history to a standstill, however, has its drawbacks. Should it not remain "the science of change," as Marc Bloch and Lucien Febvre defined it? (L'hm 250). Structuralism, which claims to be the most scientific approach, cannot fulfill this ambition. Foucault focused his attention on what is unchanging in a period or "episteme." In his system change becomes only possible "through ruptures, disjunctions and discontinuities so extreme that they effectively isolate the epochs from one another," as Hayden White noted. 18 Yet, any close inspection of the cultural products of any period shows that change is partial and gradual.

No one expects history nowadays to be deterministic in the old-fashioned way, but a rational comprehension of a chain of occurrences may be gained by a "regressive progression." The concatenation can become intelligible. Yet the historian cannot reduce all contingencies when "retracing forward what he has traced backwards." 19 Unlike the physicist he cannot claim the occurrences were predictable, but can only make them appear probable. The introduction of quantitative methods, however useful, should not create an illusion of perfect scientificity. To [End Page 532] give one instance only, taken from historical demography, the technique of "back projection" used by Peter Laslett, Anthony Wrigley, and Schofield is now thought pertinent only for a few decades, not for several centuries. 20 Le Goff and Duby both reject the incautious claim of François Furet, apparently endorsed by LeRoy-Ladurie: "Scientifically speaking, there is no social history unless it is quantitative" ("il n'est d'histoire sociale que quantitative") (L'hm 185). The worst fault of scientificity would be to ignore man, and a title chosen by LeRoy-Ladurie is disquieting: "L'histoire sans les hommes." 21 Why study the evolution of climate for nine centuries and disregard its possible influence on man and society, an influence assumed to be insignificant? It may be styled a Copernican revolution but it robs history of its meaning for us. Ricoeur warns the historian against "the fascination of a false objectivity in a history only made up of structures, forces, institutions, without the presence of men and human values" (Hv 52). History also tends to lose its meaning when it is excessively fragmented, as may happen in the so-called "serial history": François Dosse justly complains about "L'histoire en miettes" (history in crumbs).

The globalization of history, no doubt, may be--and has long been--warped by reductive ideologies and their hasty generalizations. Yet Lucien Febvre for the Renaissance, Georges Duby for the Middle Ages, Jean-Pierre Vernant for Greek antiquity have shown how global history should be conceived. "Society is a whole," Duby insists, "I do not think it possible to dissociate politics from economics or culture. Their coalescence compels us to gather informations of all sorts." 22 So does Vernant claim that Greek religion can only be understood by "thinking together the political and the religious, ethics, and daily life." 23

Literary history should also aim at being comprehensive since the evolution of literature is more or less closely related to the social, economic, and cultural context. Interart analogies must be explored, with due allowance for the differences implied by different media. In the tracing of global change the literary historian meets with the same problems as the historian of society. He can discern correlations and may even establish probable causal relationships as long as the general trends of a period are the object of inquiry. But the genesis of individual works, like the individual actions of men, is governed by a Heisenberg-like uncertainty principle. A probabilist explanation of the prevailing themes and forms of expression in a given environment at a stated time is generally possible, but the particular characteristics of each work, like the course of each elementary particle, will always partly, though not wholly, evade our grasp, as I shall further argue. But the claims of linguistics must first be considered.

Before the emergence of modern linguistics, philology proceeded [End Page 533] along lines of historical inquiry and successfully aimed in the field of language at the type of diachronic reconstruction achieved by historians in other fields. The ambition of linguistics from Saussure to the present day has been to treat language as a synchronic system of signs which has to be separated from the extralinguistic reality in order to become an object of science. The grammarian is assumed to be able to predict the way language will operate in such and such a case once the adequate instruments of analysis are discovered; prediction would then be a proof of the correctness of the analysis. 24 The syntactic system is, indeed, a structural pattern to which the extralinguistic reality must submit and the enunciator does construct his representation of the real by means of his internal grammar. Yet various languages seem to impose different patterns on this representation. If the patterning in each language were quite arbitrary, could a science of signs ever attain the universality of physical science? Whether universal laws are within the reach of linguistics is a problem I must leave to experts. I note that some languages have only past-present forms, or present-future forms, yet their speakers, when they are taught European languages, use their tense forms correctly. Universality seems to be found not in the structures of language, but in structures of the mind prior to and independent of all linguistic systems.

Certainty may be within the reach of the science of linguistics in its proper domain, and the impact of linguistic methods of structural analysis on the social sciences and the study of literature is too well known to require illustration. This expansion, however, is open to debate. As early as 1967 I had called in question the analysis of the Oedipus myth by Lévi-Strauss, pointing out the arbitrariness of the method, both in the exclusion of parts of the myth (Oedipus blinding himself) and in the inclusion of disparate facts in a single category for the sake of parallelism. 25 I was later pleased to read other objections raised by Ricoeur. 26 He admitted that the structural analysis of Lévi-Strauss had "explained the myth"--which is more than I am inclined to grant--but added that it had not "interpreted it" (IT 84), reading the text only as a text, thanks to the suspension of its meaning. To explain a text by "suspending its meaning" is unrewarding, to my mind.

Of late the reactions against the linguistic, logical, and purely formal approach to the human sciences and literature have intensified. I picked out two relevant protests against the "empire of signs," one by a poet and essayist, Yves Bonnefoy, and one by a philosopher with an interest in science, Michel Serres. The poet modestly refrains from challenging the modern episteme. He seems ready to grant that "the signified is not anchored in the referent," 27 that truth is no longer felt as the adequation of linguistic signs to a world endowed with existence, that the structures [End Page 534] we discern in the emergence of phenomena are only the consequence of the categories of language we use (254). Yet the poet is still bound to see in the word "a material, natural body," "born as breath, using our breath which makes it live outside ourselves." Through the material, natural body of the word, our own body meets the world. Saussure chose the word arbor to draw his distinctions between the signifier, the signified, and the referent. Bonnefoy probably had this instance in mind when he wrote: "if the word tree, for instance, may easily lend itself to conceptual thought, it will nevertheless, when uttered aloud in isolation, evoke the immediate and full presence of the tree, that excess always found in the very object over any form of knowledge" (262). Poetry therefore may be identified with a search for "presence" (the "presence" Derrida looks upon as an illusion), and a search for "meaning," for, "despite the strictures of the semiologists, poetry may pretend to truth" (263-64).

The philosopher speaks more harshly. "Everywhere around us," growls Michel Serres, "language is substituted for experience, the softness of the sign for the hardness of the thing. I cannot look upon this substitution as an equivalence, rather as an abuse and a violence: the tinkling of a coin is not worth the coin, the flavour of a kitchen does not fill a hungry stomach, publicity is not equivalent to quality, the speaking tongue annuls the tasting tongue, or the tongue that gives and receives a kiss." 28 Speaking of the "philosophical grammar of our age," he will concede: "it is enlightening, no doubt, but at the price of a technical language, tight, limited, turning often to algorithm, soon out of the reach of the layman's mind, as the language of scholasticism formerly was, as if the School wished to keep away those who are not equipped to take part in the conversation. Communication theory in its very rigour becomes incommunicable through an excess of technicality." 29

This, some will say, is only an expression of temper, not a refutation on a philosophical and scientific basis. I agree and think it necessary to argue: the mind is not satisfied by a divorce between the intellect and the claims of the imagination. What is proved upon our pulses should also be amenable to proof by reason. To Lacan and his followers "it is the world of words that creates the world of things--things originally confused in the hic et nunc of the all in the process of coming into being--by giving its concrete being to their essence, and its ubiquity to what has always been." 30 This is what I cannot accept. When the painter says in Browning's poem, "Fra Lippo Lippi": "This world's no blot for us, / Nor blank--it means intensely," "meaning" of course takes on a particular sense--meaningfulness--and it is projected onto the world by the painter's vision. 31 Yet I cannot believe that "all the mighty world of eye and ear," though it exists for us, as Wordsworth's phrasing admirably [End Page 535] admits, as perceived by the human eye and the human ear, is only in itself a blank, or a meaningless confusion. 32 This may be still an imaginative rather than a scientific reason for investing the referent with meaning. But one may also invoke Ricoeur's remarks on the "preverbal character" of our experience of myth and symbol and the capacity of the cosmos to signify without words: "the revealing grounds the saying, not the reverse" (IT 61, 63). Besides, as David Farley-Hills observed, "recent discussion of language tends to reject the idea of language as a closed and opaque system, seeing it more readily as dependent on our perception of the world about us and only explicable in terms of the world it tries to reflect rather than in terms of its own internal structures." 33 An elaborate counterblast to the theses of Derrida and Lacan may also be found in Raymond Tallis's Not Saussure. Tallis offers a middle way which I find attractive "between implausible alternatives: either that language is totally responsible for the differentiation of reality; or that reality is differentiated completely independently of the manner in which it is spoken of." 34 As early as 1976 Ricoeur had stated: "Language is not a world of its own. It is not even a world. But because we are in the world . . . we have experience to bring to language" (IT 20-21).

Linguistic theory did prove helpful in many fields of knowledge, but linguistic imperialism, like all empires, is probably bound to have its decline and fall. What perspectives, then, remain open in the study of literature if we look for certainty? I need not review the traditional approaches for the authentication and edition of texts: the methods of scholarship have been tested. Disagreement between experts will still arise but the principles are seldom questioned. From my present point of view the main problems concern validity in interpretation and the various ways of accounting for the genesis of a literary text.

The first stage in interpretation is the elucidation of meaning. The first problem: what kind of meaning? In Umberto Eco's phrasing we may seek what the author meant-- intentio auctoris --or what the text says in accordance with its own contextual coherence and the systems of signification to which it refers-- intentio operis --or rather seek in the text what the reader finds in accordance with his own systems of signification and his own desires and pulsions-- intentio lectoris. 35

Has the author's meaning priority? Should we even speak of the author? Roland Barthes maintained that linguistically the author is never more than the instance writing, just as I is nothing other than the instance saying "I." This claim was supported by Foucault and Derrida. I must admit that my own reaction to the alleged "death of the author" is partly emotional. For centuries readers have held a silent communion with men and women of the past through their writings: Sophocles and Sappho, Lucretius and Virgil, Chaucer and Montaigne or Marguerite de [End Page 536] Navarre, to mention only writers distant from us. Distant yet present, for when we feel in their words the presence of a mind and heart, who would turn those words throbbing with life into dead signs? Am I clinging to an illusion? If so, it was shared by Georges Poulet, and Paul Ricoeur also assumes that the written signs convey another mind's experience to us, though indirectly, that "empathy" enables us to grasp the author's meaning.

The death of the author is one of three catchwords of modernity, but are they more than rhetorical flourishes? The death of God was first announced: in our days He seems to be alive with a vengeance and the more urgent need may not be for more faith but more tolerance. The death of man was next proclaimed; but no more was meant by it than the waning of a conception of man as the subject of knowledge which had come into being with the philosophy of Kant. "Man," Foucault solemnly declared, "is only a recent invention, a figure not yet two centuries old . . . he will disappear again." 36 If this were so, I would not find it distressing to live as Montaigne and Shakespeare lived before the alleged appearance of "man." As to the death of the author, a recent book, The Death and Return of the Author, dispenses me from further comment. What Sean Burke luminously demonstrates is that Barthes himself, in fact, only objected to the author's presence in the texts that speak the language of representation: "If a text has been 'unglued' of its referentiality, its author need not die; to the contrary, he can flourish, become an object of biographical pleasure, perhaps even a founder of language" like Sade, Fourier, or Loyola. 37

To restore the presence of the author does not imply, of course, that what is known or presumed about his intention should dictate our interpretation of the text. In most cases the intention has not been stated; when stated it usually concerns only the general purport of the work. As early as Wimsatt and Beardsley's essay on "The Intentional Fallacy" in 1946 many "New Critics" had decided not to step outside the text in search of an author and Derrida only turned a method into an ontological assumption when he asserted: "There is nothing outside the text." 38 Plain minds who refuse this exaltation of the text against a background of nothingness may yet agree that texts have been enriched with new meanings by different readers in different ages. This is mere evidence, and need hardly be developed into a theory of reception. But it is widely different from the assumption that any text has unlimited possibilities of meaning. Derrida's denial of any determinate meaning derives from Gadamer's description of a text in Wahrheit und Methode as "a never-ending array of possible meanings lying in wait for a never-ending array of interpreters." 39 This view is strongly qualified by Umberto Eco, who, as the author of The Open Work, can hardly be suspected of [End Page 537] favoring fixity in interpretation. No one, he argues in The Limits of Interpretation, can maintain that all possible meanings are equally satisfactory (Li 67): all interpretations are not allowable, even for Finnegans Wake (Li 129). Admitting the difficulty of proving any interpretation right, Eco claims that it is easier to recognize wrong interpretations (Li 130).

Interpretation therefore may rest content with probability, and even delight in diversity, for interpretation is bound to remain an art. Explanation, however, should preserve the possibility of a scientific approach and at least aim at certainty.

One way of explaining is to differentiate and to categorize. Much has been achieved in that way, notably in the study of genres and of the structures of narrative and drama. New methods of analysis have sharpened our comprehension of the literary artifact and its impact upon the reader or the audience. To make the characteristics of a literary work intelligible, however, does not account for their origin.

Genetic explanations are of two kinds, social or individual. Against a strict social determinism Etienne Gilson's argument is still decisive: in the history of Western culture Plato and Aristotle stand for two main trends and two different types of mind, yet they lived in the same age and the same environment. Lucien Goldman, following Lukacs, refined the Marxist theory and agreed that the Zeitgeist was a myth. In Le Dieu caché, by narrowing the historical context to a social group with a unified structure and a highly concentrated ideology, the Jansenist group, he was able to discover analogies between various individual expressions of a common world vision, which is not surprising. It did not justify, however, his assertion that "Racine, had he lived in a different milieu, might have written plays of the same type as those of Corneille and Molière"--a hypothesis which can neither be proved nor disproved. 40

The Anglo-Saxon "new historicists" acknowledge the complexities of history and have greatly enriched our knowledge of it. But, under the influence of deconstruction, some of them tend to interpret any open statement as a means of expressing its opposite, a method far too systematic. Besides, new historicism usually negates the individual. Though fully aware of the diversity of individual response to the social context in his brilliant study of Renaissance Self-Fashioning Greenblatt himself concludes he has not discovered "an epiphany of identity freely chosen but a cultural artifact." 41

Even if we leave it to philosophers to decide whether the self has any substantive reality or not, the pragmatic observation of behavior shows that individuals obey firmly constant impulses which social determinants cannot fully explain. Personality studies or characterology as practiced by Heymans and Wiersma, Le Senne, H. J. Eysenck, and so on, have not [End Page 538] yet, and may never have, the status of exact sciences though they increasingly stress the concepts of system or structure, but they cannot be simply ignored by the philosopher and the literary critic. 42 Character is fact, if not fate.

Anthropological criticism takes us beyond society, but also beyond the individual since it creates categories of experience basic to the human psyche. It has proved immensely suggestive and fruitful. Yet the application of ready-made significations to literary works as hermeneutic devices may lead to reading in each work what has been projected upon it.

Freudian psychoanalysis focuses its attention more closely on the individual, while relating his personal experience to laws of wider application. Freud's own approach was scientific and such scientists and historians as David Ruelle and LeRoy-Ladurie, though distrusting Lacan's version, still think that psychoanalysis could be used in conjunction with neurophysiology and the techniques of history. 43 In the application of psychoanalysis to literature the trouble is not that the theory does not work, but that it always works. Starobinski also noted it: just as the Marxist critic may at once consider the literary work as a reflection of the prevailing social relations and yet account for discordant elements through the notion of mystification, so the psychoanalyst through the notions of repression and displacement can interpret any sign both
ways to suit his explanation. 44 Charles Mauron in his "psycho-critique" had wisely limited his ambition to the discovery of obsessive images within the text, disclaiming any attempt at "interpreting and explaining a whole work by reducing its complexity to some magic common denominator." 45

Since I have criticized several theories, I should lay myself open to criticism though I have to put my views in a nutshell. In my three-volume study of eight metaphysical poets, published in 1960, rather than start from an established theoretical model, I chose a purely empirical approach to genetic problems, but close investigation led to general conclusions. The influence of the social context was acknowledged, but so were its limitations. The social structures in a given environment appeared to be responsible for the statistical predominance of some characteristics in the genres and modes of literary expression chosen by a certain number of individuals, though not by all the individuals belonging to the same circle. Pride of place was therefore given to a study of the particular forms or structures of individual expression (PM I.71-75, III.9-11, 396-400). From an exhaustive analysis of the forms of perception and imagination, the modes of thought and sensibility discernible in the verse and prose of each author, a number of [End Page 539] correspondences emerged. 46 They were grouped within three systems: (1) the forms of perception of time, space, and motion, and some of the forms of thought; (2) the various modes of self-consciousness and their consequences for the affective life; and (3) a predisposition of the individual mind to grasp the sensible and the spiritual, the natural and the transnatural jointly or separately. The correlated traits were more abstract and formal than the behavioral patterns of personality studies (often of slender interest for literary analysis), yet more precise and diverse than the four kinds of "material" imagination discerned by Bachelard (PM I.53-69).

The consistency and constancy of these structures proved remarkable and, in a comparison of the different poets, it appeared that structures alike at one point tended to be alike at all other points; and that structures different at one point proved different at all other points. Though the imagination of each writer seems to be channeled by preestablished structures, this kind of inner coherence does not imply a strict psychological determinism: it leaves room for a freedom of orientation and interpretation. To give one instance, when Donne decided he would no longer make love to woman, but make love to God, he made a choice, and (to invert Milton's phrase), choice is freedom, but his conversion did not alter his fundamental modes of imagination and sensibility (PM I.68; Livre I).

When we consider the concrete contents of an individual consciousness the self is found to be plural, unstable, changing, as it appeared to Montaigne and to all later analysts of the inner life. 47 Yet some forms seem to be imposed upon individual experience by the structures of each mind, whether inborn or acquired at an early age. This kind of constancy was fully explored in the three-volume study, and later--all too briefly--in an essay on "Milton's unchanging mind" which discovered in his early poems the same psychological and formal structures as in his later and greater works despite all changes in interest, opinions, and style. 48

This is not a theory, but an observation of facts limited to a number of individuals. To assume that this constancy can be observed in all individuals is to frame a hypothesis which cannot be verified in the same way as a scientific theory since one cannot experiment. By multiplying tests on different individuals and selecting those whose lives and works show a distinct evolution, one might claim at least some probability for a general law, but the existence of such a law could be disproved, as all scientific theories may be, by a contradiction in a single case, unless the contradiction could be accounted for by another hypothesis.

To prove that a world of apparent chaos is amenable to laws, whether certain or probable, the scientist is always ready to discard earlier [End Page 540] theories and he nowadays admits that he will never pry into "the mystery of things," ultimate reality. With equal humility, the literary critic, studying the traces of the literary imagination, may at times discover whither and why its footsteps followed a preordained course, but he knows--or should know--that the living energy of the mind will always evade a formal analysis even though the creative act is revived at each new reading through the power of empathy and the "deep power of joy" (TA 49).

Sorbonne Nouvelle

Robert Ellrodt is emeritus Professor of English at the University of the Sorbonne Nouvelle (Paris III) and former President of this University. His recent publications include John Donne: Poésie (1993), and Milton: Le Paradis Perdu (1995). He has edited Genèse de la Conscience Moderne (1988) and Renaissances Européenes et Renaissance Française (1995). He is working on Doctor Faustus and King Lear for Pléiade editions of Elizabethan drama.

Notes

1. This view is qualified in an article of N. Katherine Hayles: "Chaos asOrderly Disorder: Shifting Ground in Contemporary Literature and Science," New Literary History, 20 (1989), 305-22. She admits differences between Derrida and science since "for deconstructionists, chaos repudiates order; for scientists, chaos makes order possible" (pp. 316-17); yet her essay seeks to prove that the new scientific paradigms, assumed to be "social constructions" (p. 311), are "isomorphic" with the assumptions embodied in a theory for which "indetermination inheres in writing's very essence" (pp. 314-16).

2. Bernard d'Espagnat,A la recherche du réel (Paris, 1979), hereafter cited in text as Rr; Yves Bonnefoy, André Lichnerowicz, and M. P. Schützenberger, Vérité poétique et vérité scientifique (Paris, 1989), hereafter cited in text as Vp; David Ruelle, Hasard et chaos (Paris, 1991), hereafter cited in text as Hc.

3. John Milton,Areopagitica, in Complete Prose Works of John Milton, vol. 2, ed. Ernest Sirluck (New Haven, 1959), p. 549.

4. Jean-François Lyotard,La Condition postmoderne (Paris, 1979), pp. 71-77.

5. A.-J. Greimas,Du Sens (Paris, 1983), pp. 130, 133.

6. See Arnoldo Momigliano, "L'historie entre la médecine et larhétorique," in Certitudes et incertitudes de l'histoire, ed. Gilbert Gadoffre (Paris, 1987), p. 34; hereafter cited in text as L'he.

7. See Hermann Weber, "Jean Bodin et la vérité historique,"in Certitudes et incertitudes, p. 79.

8. See Marc Fumaroli, "Historiographic et épistémologieà l'époque classique," in Certitudes et incertitudes, pp. 90-96.

9. Hayden White, "The Historical Text as Literary Artifice,"CLIO, 3 (1974), 227-303.

10. Roland Barthes, "La discours de l'histoire," inSocial Science Information, 6 (1967), 65-75; Michel de Certeau, L'écriture de l'Histoire (Paris, 1984), p. 54.

11. Hans-Robert Jauss, "Experience historique et fiction," inCertitudes et incertitudes, pp. 117-18.

12. The "fictional" view of history is also sharply rejected by F.Bédarida and Etienne François, in Passés recomposés: Champs et chantiers de l'histoire, ed. Jean Boutier and Dominique Julia (Paris, 1994).

13. Paul Ricoeur,Histoire et vérité (Paris, 1955), p. 28; hereafter cited in text as Hv.

14. Carl G. Hempel, "The Function of General Laws in History,"Journal of Philosophy, 39 (1942), 35.

15. Paul Ricoeur,Temps et récit (Paris, 1983-85), I.161-65; hereafter cited in text as Tr.

16. Philippe Ariès,Un historien du dimanche (Paris, 1980). On Paul Veyne, see Ricoeur, Temps et récit, I.239-44.

17. See François Dosse,L'histoire en miettes (Paris, 1987), pp. 130ff.; hereafter cited in text as L'hm.

18. Hayden White,Tropics of Discourse (Baltimore, 1978), p. 235; quoted by Richard Lehan, in "The Theoretical Limits of the New Historicism," New Literary History, 21 (1990), 539.

19. As Louis Mink admits in "Philosophical Analysis and HistoricalUnderstanding," Review of Metaphysics, 21 (1968), 222.

20. LeRoy-Ladurie, "Quelques orientations de la Nouvelle Historie," inCertitudes et incertitudes, p. 152.

21. LeRoy-Ladurie, "Quatrième partie," inLe territoire de l'historien (Paris, 1973).

22. Georges Duby,Dialectiques, 10-11 (1975), 121.

23. Jean-Pierre Vernant,Nouvel Observateur, no. 808 (5 May 1980), 54-62.

24. Henri Adamczewski, "Extra--et métalinguistique: La langueà la rencontre du réel," RANAM, 19 (1986), 26.

25. As in the illustration of the overrating of blood relations by suchincidents as Cadmos seeking his sister, Oedipus wedding his mother, Antigone burying her brother. See my Les fins et les fondements de la critique (Montpellier, 1969).

26. Paul Ricoeur,Interpretation Theory (Fort Worth, Tex., 1976); hereafter cited in text as IT.

27. Yves Bonnefoy,Entretiens sur la poésie (Paris, 1990), p. 286; hereafter cited in text.

28. Michel Serres,Eclaircissements (Paris, 1992), p. 193.

29. Michel Serres,Le Tiers-Instruit (Paris, 1991), p. 125.

30. Jacques Lacan,Ecrits I (Paris, 1966), p. 155. It is fair to quote the whole sentence, not the first clause only: Lacan does not deny the existence of an extra-linguistic reality (nor does Derrida). But he seems inconsistent when he speaks of the "essence" of the things "confused" as if they had a specific being before being named and yet maintains that "the concept engenders the thing" and "is the thing itself," that the word "is the trace of a nothing" (p. 155).

31. Robert Browning,Poetical Works: 1833-1864, ed. Ian Jack (London, 1970), ll. 313-14.

32. William Wordsworth, "Tintern Abbey," inThe Poetical Works of Wordsworth, ed. Paul D. Sheats (Boston, 1982), ll.105-6; hereafter cited in text as TA by line.

33. David Farley-Hills, "Deconstruction: A Deconstruction,"Essays in Criticism (July 1992), 181.

34. Raymond Tallis,Not Saussure (Macmillan, 1988), p. 124.

35. Umberto Eco,Les limites de l'interprétation (Paris, 1990), pp. 29-30; hereafter cited in text as Li.

36. Michel Foucault,The Order of Things, tr. A. Sheridan (London, 1970), p. xxiii.

37. Sean Burke,The Death and Return of the Author: Criticism and Subjectivity in Barthes, Foucault and Derrida (Edinburgh, 1992), p. 47.

38. Jacques Derrida,Of Grammatology, tr. Gayatri C. Spivak (Baltimore, 1976), p. 158.

39. Hans-Georg Gadamer,Wahrheit und Methode (Tübingen, 1960), quoted by E. D. Hirsch in Validity in Interpretation (New Haven, 1967), p. 249.

40. Lucien Goldman,Le Dieu caché (Paris, 1958), ch. 5.

41. Stephen Greenblatt,Renaissance Self-Fashioning: From More to Shakespeare (Chicago, 1980), p. 256.

42. See H. J. Eysenck,The Structure of Human Personality (London, 1953), René Le Senne, Traité de caractérologie (Paris, 1945), and my discussion of its application to literature, Poètes métaphysiques anglais, 3 vols. (Paris, 1960), I.58-69; hereafter cited in text as Pm by volume and page.

43. David Ruelle,Hasard et Chaos, p. 206; LeRoy-Ladurie, "Quelques orientations."

44. See Jean Starobinski, "Considerations of the Present State ofLiterary Criticism," Diogenes, 74 (1971), 74, 79ff.

45. Charles Mauron,Des métaphores obsédantes au mythe personnel (Paris, 1964), p. 13.

46. A table of correspondences is given in II, pp. 411-22.

47. See my essay, "L'interrogation sur l'identité de Montaigneà Donne," in Renaissances européennes et Renaissance française, ed. Gilbert Gadoffre (Montpellier, 1995).

48. Robert Ellrodt, "Milton's Unchanging Mind and the Early Poems,"Milton Quarterly, 22 (1988), 59-62.