4. Reflections on the Practice of Philology*


To edit a page of Aeschylus or Plato, to separate these authors’ sentences or their lines of verse—to understand, in a word—is to discover that, to a considerable degree, the composition obtained by a lengthy and coherent tradition, but also by modern science, heir to that tradition, is different from the text that has been preserved, sometimes only in an interrupted fashion within the critical apparatus at the bottom of the pages, but also on the upper part of the pages, in the lines of a standard reading. To go back to even a single well-known sentence is often to reconstruct it according to a different grammar. This doubling of the text at the literal level requires an explanation. It cannot be the product of error, since every analysis stems from precise conditions that justify it, nor can it result from a taste for contradiction. Still, the phenomenon invites us to meditate on the frontiers of hermeticism and the very possibilities of communication.
Several books on classical authors, books that are also editions of these authors’ works, have led me to reflect on the reason for the difficulties inherent in the practice of philology. To establish and translate canonical texts was to encounter, from paragraph to paragraph, the protean personality of the interpreter, fettered by authority, and to observe academic ways of proceeding. Most readers are unaware that they are reading Marullus in Lucretius, Usener in Epicurus, Diels in Heraclitus. The text defends itself poorly. It is weak in the face of conjecture. For there were all sorts of good reasons to impose corrections, reasons that are still prevalent today, because biases are a constant and because tradition has the merit of existing and has the force of established fact.
Examples in which description is distorted in an analogous way could be found in abundance among the most widely-read authors. And yet all the flourishes of free, “creative” reading are poorer than the program that would be imposed by the systematic establishment of the unrecognized dimension in question. Elements that are proper to a text, instead of being analyzed as signs, are characterized as accidental, interpreted as instances of discord or incoherence. The integration of such terms shows on the contrary that the logic of the text is precisely not the logic that has caused these elements to be viewed as discordant. A given sentence no doubt conceals both truth and error, and the creative force it possesses would presumably be able even to trace the route of phantom navigators across the sea. The Ancients (for example, Diogenes Laertius) said that there was a Heraclitus for fools, and probably a Homer as well. Certain interpretations settle for dissociating what the genesis of the work had melded together, or else they reassemble what the genesis of the work had taken apart. [1]
It would be easy—and unfair—to scoff at the efforts expended in the ever so cumbersome apparatus of philological demonstration. Immediate access to the text is demanded insistently. And yet it is impossible to build an interpretation based on the raw material without going through the history of interpretations. Not that the centuries have supplied progressively more accurate approximations on the path toward truth, nor that it is profitable a priori to draw the components of a new mixture from the multiple elements encountered; but, beyond the fact that the justification of the material—that is, of the text retained—cannot be achieved unless one recognizes and refutes the reasons for which it has been altered, the material itself is constituted only to the extent that one interpretation imposes itself as opposed to others (the hermeneutic circle also applies to the discovery of the text over time). No reading of any sort can become clear in the absence of other readings. A correct reading asserts itself only when it is confronted, on its own, with all possible readings.
Even if this heuristic aspect did not exist, a demonstration would not be convincing if it did not win out through refutation. It would appear, owing solely to the fact of its difference, to be more esoteric than the text explained. We know that it does not suffice to argue in order to convince. We know that, in a culture that has always been based on the explication of texts, the greatest share of the commentary produced has no justification but the smooth functioning of the institution—wherever that institution may be—and the perpetuation of the exercise. The quantity, the mass of publications, then forms an insurmountable obstacle, and the decline in real interest has not led to a decrease in production.
Moreover, technique, in the discipline of philology as in others, has a tendency to become automatic with respect to its object. The texts then enter into the patrimony of literary culture, which is an established and capitalizable value. The division of labor, problematic in this domain and difficult to reconcile with the distinction between technicians and users, nevertheless takes its legitimacy from the virtual number of workers. A good many technicians, and among the most productive, like to say that they understand nothing, not only about the “ultimate signified” but about meaning of any sort; working blind, they fabricate the material that others, the technicians of meaning, receive from them, fully prepared. This divorce casts suspicion on all the histories of philosophy and on most of the chapters in the history of the sciences. The unfounded constructions in these histories, suspended over uncertain material, weigh in turn on interpretations, that is, on the text in question.
In Empedocles there is a theory, fully developed but wholly consigned to oblivion, according to which the sun that we see is produced in the vault of heaven by the reflection of light on the earth. Not only does this theory explain the fragments of sentences that remain to us, but it is richer than a simple opinion on a point of science would be. It allows us to see the positions of all the bodies in the universe, which are always in relation with the theory. [2] Similarly, the system of blue eyes and black eyes is in its own right the complete, encoded expression of a theory of vision. [3] Now, these two components, discovered in the thickness of a doxographic summary, are almost irremediably lacking in histories of astronomy or optics, because their place in a given constellation did not allow them to enter easily into a linear history of scientific progress, and this progress is necessarily the guiding thread of virtually all particular histories.
The division of labor is probably inevitable in practice, but it is unacceptable in theory, and it is maintained only by self-interest. [4] If the history of philology is to destroy this recourse to partial legitimacies, it can no longer be a history of humanists and learned scholars, [5] or even a history of the perfecting of techniques, but rather must become an analysis in reverse, one that would go back through time to situate interpretations, or more precisely their historical determinations. Not that one would thereby arrive at the meaning of a given text or even of a particular word, and here is the heart of the conflict: it lies in the difference that separates the object that is being pursued in an interminable doxography from the object that properly belongs to the science of discourse. At least in this way, philology, being a mirror of non-comprehension, becomes an invaluable way of studying societies. And one might want to show, in the years to come, that what the history of philology and interpretations brings to light applies just as much to contemporary cultural systems. Even today, translations are still permeated by national traditions.


At the end of the nineteenth century, the Germanic philological-historical method found itself, in the face of its object, in a state of radical contradiction. Reflection on the constitution of knowledge—in other words, traditional hermeneutics—was reduced to its simplest expression, to the point of being merely an appendix to “formal” philology, which was conceived for its part according to the model of the philology of the real (Realphilologie). [6] Philology, destined at that time chiefly to smooth over and correct, took pride in its role as censor rather than bemoaning censorship as a thankless task: “It is the role of formal philosophy to follow the expression of the thought, which is sometimes deficient, and to a certain extent to monitor the creative work.” [7] The analyst of the written expression steps in as psychologist and therapist at the end of the process of exploring meaning.
That the fundamental science we know as exegesis should be degraded in this way, to the point of holding itself back and censoring itself as it censors the text, in order to devote itself exclusively to the hothouse of correction or suppression, signifies that philology is riven by an internal contradiction. Censorship, paradoxically, defends values that it does not consider its own. Setting itself up as a science, and as such—in order to gain recognition as such—viewing all its objects as equivalent, it ultimately treats authors in a mercenary fashion, since it leaves to the vague and deplored field of aesthetics the responsibility for protecting their reputations. It masks absence of judgment by the amplitude of its scientific apparatus. Subsequently, literary theory has come to occupy the space, without including the conditions under which meaning is produced. The gulf has been deepened. Material difficulty is taken to be more serious than intellectual difficulty. “If Euripides’ poems were not corrupted,” Wilamowitz writes in his “Introduction to Greek Tragedy,” “Hermann would not publish them.” [8]
Philological censorship is in power wherever it is charged with a pedagogical mission, but it uses its weapons far from the pedagogical context in order to bring about a rapid increase in the material of knowledge—and the frenetic rhythm of fabrication is not the smallest failing in the activity favored by the philological-historical method. On this basis, classical antiquity no longer has any raison d’être, even if philology scrupulously carries out its mission, which is to be at the service of classical antiquity.
The contradiction has been manifested both in individual critics, capable of holding two different discourses depending on whether they are representing the science of philology from an academic platform or praising the values that allowed them to practice that science, [9] and also in the war waged between official science and the inner circles that, in reaction, celebrated the offices of aestheticism. On the Homeric question, for example, the quarrel between the Analysts and the Unitarians is situated almost entirely in this necessary tension between dignitaries and marginal figures. The former are invested with the authority of the academy; the latter are justified by faith. Thus scholars have never been able to get beyond the antinomy through peaceful examination, through a scientifically-based Unitarianism.
The perfecting of tools during the nineteenth century—the inventory and classification of manuscripts, the history of texts, and paleography—offered the support of material accidents as a way of dealing with incomprehension. Thus lacunae and interpolations flourished from a certain point on: at a given moment, between 1850 and 1860, the editions produced by Bernays, Lachmann, Munro, and Brieger broke Lucretius apart by more lacunae than had ever been recognized before, to the point that one of these editors expressed his own amazement at the richness of his harvest. One can take all the so-called interpolations of a tragedy of Aeschylus, as I have done with Pierre Judet de La Combe for the Agamemnon, for example, in order to show that all of them have been produced by a sometimes traditional difficulty in comprehension which, for various reasons, had never found another solution. [10] The contemporary tendency to give up, ready to condemn the earlier excesses, changes nothing, because it does not see the link between these interpolations and the practice that consists in resolving difficulties of meaning by invoking accident. In order to be brought back to life, the lost “meaning” waits to be integrated into any explanatory system whatsoever.
Although the appeal to systems of representation and to mental and psychological structures supplies interpretations and, above all, grids, such studies have little to do with real knowledge of the works themselves. It is true that their authors are searching for an illuminating and all-embracing hermeneutic meaning, but the light that is borrowed directly from social reality—generational conflict, polis, genos, and so on—makes this reality a purely abstract matter, too quickly enlarged to the dimensions of a culture or an epoch, and it makes the play or poem a mirror of the schema retained, which is imposed in all cases, ipso facto. Yet the very self-evidence of the example is suspect. Because they fail to take an interest at the same time in the unique meaning of the work, the essayists of the structuralist schools do not really use the traditional tools of textual criticism and thus do not succeed in critiquing the inherited meanings.
To be sure, particularity of expression has been recognized since Wilamowitz’s day as a proper historical form, and it has been related not only to systems of representation but also to types of teaching and to rhetorical habits.
Reference to the humanist norm has been abandoned among the formalists in favor of specific modes of expression, surely under the influence of ethnology and anthropology. The alienation of the author’s viewpoint is perceptible in these writings. Beyond the fact that, in the case of writers such as Pindar, [11] one might ask whether the author is explained by the manual or the manual by the author, this normalization to the second degree precludes, in a schematic assimilation, a deep understanding of the properly esoteric dimension and of the level of “allegory.” Only certain simplifications—biographical references, for example—are sacrificed in relation to all that was previously being done in more psychological critiques.


The constitution of a horizon of representations is no doubt indispensable, but it must be achieved along different pathways. Wedded from the outset to the idea of an alterity of material and the possibility of historical objectivization, not so much practicing the naturalization of the foreign as postulating its existence, historians project contemporary biases, if only by applying models from all over the place. There is nothing more historically determinable than the introduction of explanatory schemas considered a priori as historically applicable.
August Boeckh, in his Encyklopädie und Methodologie der philologischen Wissenschaft (Encyclopedia and Methodology of the Science of Philology, the product of a course Boeckh taught regularly at the University of Berlin between 1809 and 1865), [12] defined philology as a historical science. He made it a universal discipline, alternating with philosophy, like Castor with Pollux, but at the same time his view of it encompassed the historical aspect that also characterized every science. Although remote, the object, for him, concerned the interpreter directly, and it was thus on the basis of what he knew that the interpreter was led to judge (“foreign ideas are not yet ideas for me,” he wrote [13] ). The universal philologist, the concept of which he was creating, someone capable of knowing everything since all the objects of history are known through him, introduces distance only in judgment, which is no longer historical when it has appropriated the foreign. Boeckh invites the philologist to tear himself away from his condition and rise above the thing integrated so as to have it in front of him in its atemporal objectivity. It is a matter of knowing what one knows, which has already been known by another. The judgment brought to bear on the critical act operates the appropriation, so that the doubled object is no longer the one that the foreign author had known but belongs to the knowing subject. Historical distance is abolished not only by the concept of a universal history, but also by the agency of judgment that appreciates or invalidates.
Boeckh’s hermeneutics—and Boeckh moreover was one of the founders of the science of history—thus reconciled the necessity of investigation and critique, on the one hand, and the obligation to maintain privileged values, values that concern us, on the other; in spite of everything, it closely resembled the hermeneutics codified by Friedrich Ast. [14] Boeckh situated himself in the lineage of Schelling’s philosophy of identity, without however believing it possible to purge objects of their historical contingency in order to grasp the movements of the unifying spirit. Other theorists of the eighteenth century and of German idealism had, on the contrary, like Hölderlin, by virtue of the paradigmatic nature of antiquity, ended up espousing the absolute alterity that made it possible for modernity to be recognized.
Ast’s theory, which affirms the possibility of identification, nevertheless makes distance a problem, whereas the later philologists, renouncing appropriation, renounce the possibility of knowing the remote object and end up assimilating it.


Determined by his own position in history and in society, the interpreter, an artisan of his material, occupies with respect to his object a place that, in the case of the works of antiquity, is situated at such a great distance that variations between years give him scarcely any sort of handhold. Such distance cannot count, it is absolute. In reality, if the only thing one knows is the unknown, it is this very tension from which one must profit. Moreover, there is just as much distance between Heraclitus and his contemporaries, or between Epicurus and his, so that foreignness in time is accompanied by difference in synchrony.
To objectivize, one would have to agree to take the thing studied as a fixed entity, to take life as fact, to take a play as an opaque event, to take a thought as a body. Now, what one grasps cannot be, in the instantaneous act of grasping, merely a movement in search of itself, merely the inflection in a sentence. It is not the common, known meaning that one is seeking, but the unknown, the unsaid that is revealing itself, combined with the letter of the text. Here is the specific object of philology, the one it shares with no other science.
Because it was seeking literal meaning, philology, in its positivist and chiefly documentary period, renounced allegory, the sensus spiritualis that it took to be gratuitous and superfluous; as a result, it did not reach the sensus litteralis to which prudence had led it to cling. This is because the relation between the two is not arbitrary, as philologists had believed. Allegory is not only the transfer of a meaning into a different body of thought, as when the Stoics take up Homer anew or the Church Fathers go back to the Old Testament. The “allegory” that transforms fixed discourse by putting it differently (allēgorein) belongs to speech. A sentence is not exhausted in what it says. There remains another sentence that is read in another author or in another part of the work; every sentence contradicts another sentence.
Hence, the preliminary necessity of extensive reading, and in the case of the lacuna-ridden information proper to antiquity, the necessity of complementary imaginative readings that decipher the reconstituted texts. The example of Heraclitus is particularly striking because, once the principle of implicit correction was discovered, Heinz Wismann and I were able, in Héraclite ou la séparation, to establish that, in their quasi-totality, the adages, no matter how brief, refuted an affirmation, refuted even the negation of an affirmation. In many cases, the two previous stages are not attested, and yet they can be extrapolated. [15]
With other authors, the allegorical practice does not refer to formulas brought forward by a book culture; it has to do with the internal components of the discourse and constitutes the thickness of the speech. This stratification has nothing to do with the use that is made today of the notion of levels of signification, which implies differentiated levels of consciousness. The virtual meaning, as all these complicated compositional embeddings show, is not to be confused with the resurgence of a repressed feeling. The phenomenon is linked rather to a mode of writing that Servius’ biography of Virgil brings to light: the poet dictated a large number of verses to a slave every day; he then reduced the number to a single one. The “allegory,” in this case, would be all the suppressed verses, which turn up elsewhere in other parts of the work.
Here I am no doubt describing, under the single term allegory, a phenomenon that has diverse applications, and no doubt when other studies have been completed it will be possible to try to establish their typology in relation to the ancient theories of obscuritas. There is a closed form in which the relationship among the elements is imposed as the matter to be deciphered, and there is an open form in which autonomous structures interpenetrate. The first corresponds to the philosophy of Being, the other to atomism. The one is obscure from the start and becomes clearer as one reads, the other is clear in appearance and becomes complicated as one reflects. And yet the richer the extraction of these deep meanings, the more it is taken to be subjective. Philology ought to have criteria capable of distinguishing between a fantasy that is taken to be a scientific interpretation and a close reading of the text that, because it astonishes, is taken to be a personal reading.
The literal meaning decides. When reading has finally been able to opt for a text, this literal meaning is not an “interpretation” on the part of the interpreter. The philologist is bound by the interpretation that is the text. This literal meaning is not just any term that could be shifted around like a pawn and set into relation with others in a formal game, for the good reason that this game is one that the author himself has already played. The “allegory” has produced the grammatical construction. The arrangement of words is as it is because it is determined by other arrangements that the position taken by the author includes.
The issue, then, is not that a sentence, bound by these internal links, refers to a single signified. But it is not conceivable that polysemy should escape the project, because the project includes and maintains it. All the variants of interpretive desire that are not positioned along the axis constituted by the project—a project that is historically locatable, even though it is situated only by interpretation—exceed the strict limits of philology and hermeneutics. The work is inexhaustible, but only on its own terms, according to its own autonomy. From Empedocles’ poem Aristotle certainly managed to draw answers that the poet had not provided. But Aristotle was right to add at once that, if someone had pointed out to him what he was doing, he would not have denied it. [16] Polyvalence has this avowal as its limit. It is undoubtedly difficult—but difficulty is what we are dealing with here—to trace the line of demarcation that separates the elements that the author uses fully from those that work for him. The teleology of the project encompasses them all equally.
Thus ambiguity must not be posited as a principle of composition, and the authors who make that mistake, in the case of Heraclitus or of authors of tragedies, miss the meaning revealed by the device they are studying. This device presupposes a thetic act before it is enriched by the very counter-meaning with which the author has charged it.
The readings that follow the work and transform it, annulling its particularities in a process of assimilation or using it as a point of departure, are new projects for the commentator. The stages of this afterlife (Wirkungsgeschichte) are of interest to the interpreter of the primal work only to the extent that they obscure its comprehension.
What is proper to the work, its difference, is another criterion. Among all the improbable meanings, one imposes itself (probatur), one that is even more unexpected. The limit, the criterion, would be, here, the entire set of meanings, provided that a given arrangement could be absolutely differentiated in its particular quality. Parallel passages can never confirm an interpretation, far from it; repetition in different contexts cannot be a restatement:
Even if—and this is doubtful in itself [as Peter Szondi writes in his analysis of a poem by Paul Celan]—one could claim that one or several expressions in the two places are identical, and that the interpretation taken as certain in one of the passages seems to clarify the meaning of the use of the same word in the line one is seeking to understand, this line becomes clear without being understood, since it is what it is only in this particular use and at first rightly withstands comprehension. [17]
If agreement with other homologous sections supplies the means for deciding, it is on condition that one knows at every moment that the agreement is composed of distinct and unheard-of properties that decide in their turn. The constitution of coherence comes about in a circular movement that may never come to an end. Finally, experimentation has the upper hand. As it strives to reproduce the movement of the work in the literal reading, it can try out all possible meanings. The only thing that would bring it to a halt would be the hermeneutic effectiveness of the meanings uncovered. The right explanation is richer than wrong ones.

Works Cited

Ast, F. 1808. Grundlinien der Grammatik, Hermeneutik und Kritik. Landshut. In French as “Élements de grammaire, d’herméneutique et de critique,” in Critique et herméneutique dans le premier romantisme allemand, ed. and trans. D. Thouard, 287–314.Villeneuve-d’Ascq, 1996.
Boeckh, A. 1966. Encyklopädie und Methodologie der philologischen Wissenschaften. Ed. E. Bratuschek. Darmstadt. Orig. pub. 1886.
Bollack, J. 1975. La pensée du plaisir. Épicure: textes moraux, commentaires. Paris.
———. 1981. L’Agamemnon d’Éschyle: le texte et ses interprétations. 3 vols. Paris.
———. 1992. Empédocle. 3 vols. Paris. Orig. pub. 1965, 1969.
Bollack, J., and H. Wismann. 1995. Héraclite ou la séparation. Paris. Orig. pub. 1972.
Bundy, E. L. 1962. Studia pindarica. Berkeley.
Gercke, A., ed. 1910. Einleitung in die Altertumswissenschaft. Leipzig.
Gudemann, A. 1967. Grundriss der Geschichte der klassischen Philologie. Darmstadt. Orig. pub 1907.
Pfeiffer, R. 1976. History of Classical Scholarship: From 1300 to 1850. Oxford.
Sandys, J. E. 1967. A History of Classical Scholarship. New York. Orig. pub. 1908.
Szondi, P. 1981. Poésie et poétique de la modernité. Ed. M. Bollack. Lille.
Wilamowitz-Moellendorff, U. von. 1959. Euripides Herakles. 3 vols. Darmstadt. Orig. pub. 1889.


[ back ] * Originally published as “Réflexions sur la pratique,” in: Jean Bollack, La Grèce de personne: les mots sous le mythe (Paris, 1997), pp. 93–103.
[ back ] 1. See Chap. 3, “Odysseus among the Philologists.”
[ back ] 2. See Bollack 1992, 2, frag. 322–339, and the commentary in Bollack 1992, 2:263–277.
[ back ] 3. Bollack 1992, 2, frag. 420, §8, and 437–438, and the commentary in Bollack 1992, 2:335–338.
[ back ] 4. On the division of the philological domain, see my preface to Bollack 1975:ix–xliii, and especially p. xxiv.
[ back ] 5. See the history of the humanists in Sandys 1967, Gudemann 1967, and Pfeiffer 1976.
[ back ] 6. On this notion, see Gercke 1910, in the chapter titled “Methodik,” 35.
[ back ] 7. Gercke 1910:35.
[ back ] 8. “Einleitung in die griechische Tragödie,” in Wilamowitz-Moellendorff 1959, 1:239.
[ back ] 9. See Chap. 3, “Odysseus among the Philologists.”
[ back ] 10. See Bollack 1981 and 1982.
[ back ] 11. See Bundy 1962.
[ back ] 12. Boeckh 1886.
[ back ] 13. Boeckh 1886:20.
[ back ] 14. Ast 1808.
[ back ] 15. Bollack and Wismann 1995.
[ back ] 16. Aristotle Metaphysics 1.10, 993a23.
[ back ] 17. “Lecture de Strette,” in Szondi 1998:165.