GIORGIO AGAMBEN - Potentialities Collected Essays in Philosophy

PART ONE
Language

§ 1 The Thing Itself
For Jacques Derrida and in memory of Giorgio Pasquali

The expression "the thing itself," to pragma auto, appears at the beginning of the so-called philosophical digression of Plato's Seventh Letter, a text whose importance for the history of Western philosophy has yet to be fully established. After Richard Bentley had come to suspect the entire Platonic corpus of letters of being fraudulent, and Christoph Meiners (in 1783) and subsequently Karsten and Friedrich Ast declared them to be inauthentic, Plato's letters--which until then had always been considered a central part of the philosopher's work--were slowly expelled from philosophical historiography, precisely when it was most fervent and active. When philological opinion began to change in our century, and more and more critics asserted the authenticity of Plato's letters (the letter that interests us is by now generally considered to be genuine), philosophers and scholars had to break the hundred-year-old quarantine of the Platonic epistles if they wanted to study them at all. What had been lost in the meantime was the living connection between text and philosophical tradition, with the result that the philosophical excursus contained in the Seventh Letter appeared as an arduous, solitary fragment resisting any attempt at comprehension. Naturally, it was also transformed by its long isolation into something rich and strange, which could be considered with a freshness probably unattainable in regard to any other Platonic text.
The scenario of the letter is well known: the seventy-five-year-old Plato tells Dion's friends of his encounters with Dionysius and the dramatic failure of the latter's Sicilian political projects. In the passage that interests us here, Plato recounts the story of his third stay in Sicily. Once again on the island because of the tyrant's persistent invitations, he decided to put Dionysius to the test concerning his professed desire to become a philosopher. "Now there is a method," Plato writes, "of testing such matters which is not ignoble but really suitable in the case of tyrants, and especially such as are crammed with borrowed doctrines; and this was certainly what had happened to Dionysius, as I perceived as soon as I arrived." 1 Men such as these, he continues, should be immediately shown the whole thing (pan to pragma) and the nature and number of its difficulties. If the listener is truly equal to "the thing," he will then think that he has heard the tale of a wonderful life, which must be led without delay and to which he must devote himself at all costs. On the other hand, those who are not truly philosophers and have only an outer glow of philosophy, like those whose skin is tanned by the sun, will see the difficulty of "the thing" and think it too hard or even impossible, convincing themselves that they already know enough and need nothing more. "This, then," Plato writes,
was what I said to Dionysius on that occasion. I did not, however, expound the matter fully, nor did Dionysius ask me to do so; for he claimed that he himself knew many of the most important doctrines and was sufficiently informed owing to the versions he had heard from his other teachers. And I am even told that he himself subsequently wrote a treatise on the subjects in which I instructed him, composing it as though it were something of his own invention and quite different from what he had heard; but of all this I know nothing. I know indeed that certain others have written about these same subjects; but what manner of men they are not even they themselves know. But thus much I can certainly declare concerning all these writers, or prospective writers, who claim to know the subjects with which I concern myself [peri ōn egō spoudazō], whether as hearers of mine or of other teachers, or from their own discoveries; it is impossible, in my judgment at least, that these men should understand anything about this subject. (Epistle VII, 341 a 7-c 4; pp. 529-31)
It is at this point that Plato uses the expression to pragma auto, the thing itself--a formulation that remained so determining as an expression of the cause of thinking and the task of philosophy that it appeared again almost two thousand years later, like a watchword passed on from Kant to Hegel, and then to Husserl and Heidegger: "There does not exist, nor will there ever exist, any treatise of mine dealing with this thing. For it does not at all admit of verbal expression like other disciplines [mathēmata], but, after one has dwelt for a long time close to the thing itself [peri to pragma auto] and in communion with it, it is suddenly brought to birth in the soul, as light that is kindled by a leaping spark; and then it nourishes itself [auto heauto ēdē trefei]" (341 c 4-d 2; p. 531).
This passage has been cited countless times as proof of esoteric interpretations of Plato and as irrefutable documentation for the existence of Plato's unwritten doctrines. According to these readings, the dialogues transmitted by our culture for centuries as a venerable legacy would not address what Plato was seriously concerned with, which would have been reserved for a purely oral tradition! This is not the place to take a position on this problem, which is surely an important one. We shall instead seek to consider the nature of the "thing itself" of which Plato speaks and which Dionysius wrongly thought he understood. What is the thing of thinking?
An answer to this question can follow only from an attentive reading of the next passage, which Plato defines as a "story and wandering" (mythos kai planos) (344 d 3; p. 541) and also as a "certain true argument, which . . . although I have frequently stated it in the past, also seems to be in need of repetition at the present time" (342 a 3-7; p. 533). Any thought that wants to grasp its "thing" must thus always reckon with interpreting this "extravagant story." Let us then attempt to read it. "Each being," Plato writes,
has three things which are the necessary means by which knowledge of that being is acquired; the knowledge itself is a fourth thing; and as a fifth one must posit the thing itself, which is knowable and truly is. First of these comes the name [onoma]; second, the definition [logos]; third, the image [eidōlon]; fourth, the knowledge. If you wish, then, to understand what I am now saying, take a single example and learn from it what applies to all. There is something called a circle [kyklos estin ti legomenon], which has for its name the word we have just mentioned; and, second, it has a definition, composed of names and verbs; for "that which is everywhere equidistant from the extremities to the center" will be the definition of that object which has for its name "round" and "spherical" and "circle." And in the third place there is that object which is portrayed and obliterated, which is shaped with a lathe and falls into decay. But none of these affections is suffered by the circle itself [autos ho kyklos, which here is the example of the thing itself], to which all these others are related, for it is different from them. The fourth is knowledge and intelligence and true opinion regarding these objects; and all this must be conceived as a single thing, which exists neither in voices [en phōnais] nor in corporeal figures [en sōmation skhēmasin], but in souls [en psychais]. Hence it is clear that it differs both from the nature of the circle itself and from the three previously mentioned. Of those four, intelligence is closest in kinship and similarity to the fifth; the others are further removed. The same is equally true of the straight figure and the sphere, color, and the good and the fair and the just, and of all bodies, whether made or naturally produced (such as fire and water and all such substances), all living creatures, and ethos in the soul and all creations [poiēmata] and passions [pathēmata]. For if someone does not grasp the first four for each thing, he will never be able to participate perfectly in knowledge of the fifth. Moreover, the first four things express the quality [ti poion ti] of each being no less than its real essence, on account of the weakness of language [dia to tōn logōn asthenes]. This is why no man of intelligence will ever venture to entrust his thoughts to language, especially if the language is unalterable, like language written with letters. (342 a 8-343 a 3; pp. 533-35)
Let us pause for a moment to catch our breath. In the face of this extraordinary excursus, which constitutes the final and most explicit presentation of the theory of the Ideas, we can measure the damage done to philosophical historiography by the nineteenth century's claim of the Platonic epistles' falsity. It is not my intention to climb that impervious massif. But it is certainly possible to seek to establish a first trail, to determine the difficulty of the climb, and to situate it with respect to the surrounding landscape.
One remark that we can make (and that has already been made by, among others, Pasquali) concerns the status of unsayability that the Seventh Letter, according to the esoteric reading of Plato, would ascribe to the thing itself. This status must be tempered by the fact that from the context it is clear that the thing itself is not something that absolutely transcends language and has nothing to do with it. Plato states in the most explicit fashion that "if the first four [which, we recall, include name and logos] are not grasped" it will never be possible fully to know the fifth. In another important passage in the letter, Plato writes that the knowledge of the thing itself suddenly emerges in "rubbing together names, definitions, visions and sense-perceptions, proving them in benevolent proofs and discussions without envy" (344 b 4-7; p. 541).
These unequivocal statements are, moreover, perfectly coherent with the very close relation between the Ideas and language that is suggested by the Platonic dialogues. When in the Phaedo Socrates presents the genesis of the Ideas, he says, "it seemed to me necessary to seek refuge in the logoi, to find the truth of beings in them" (99 e 4-6). Elsewhere, he presents the hatred of language as the worst of evils ( Phaedo, 89 d 2) and the disappearance of language as the loss of philosophy itself ( Sophist, 260 a 6-7); in the Parmenides, the Ideas are defined as "what can be apprehended to the greatest degree by means of logos" (153 e 3). And does not Aristotle, in his historical reconstruction of Plato's thought at the beginning of the Metaphysics, state that the theory of Ideas was born from a skepsis en tois logois, a search in language (987 b 33)?
The thing itself therefore has its essential place in language, even if language is certainly not adequate to it, on account, Plato says, of what is weak in language. One could say, with an apparent paradox, that the thing itself, while in some way transcending language, is nevertheless possible only in language and by virtue of language: precisely the thing of language. When Plato says that what he is concerned with is in no way sayable like othermathēmata, it is therefore necessary to place the accent on the last three words: it is not sayable in the same way as other disciplines, but it is not for that reason simply unsayable. As Plato does not tire of repeating (341 e 1-5), the reasons why it is inadvisable to entrust the thing itself to writing are ethical and not merely logical. Platonic mysticism--if such a mysticism exists--is, like all authentic mysticism, profoundly implicated in the logoi.
Now that we have made these preliminary observations, let us closely examine the list contained in the digression. The identification of the first four members does not pose any great difficulties: name, defining discourse, image (which indicates the sensible object), and, finally, the knowledge achieved through them. Name (onoma) is, in modern terms, which are those of Stoic logic, the "signifier"; logos is the "signified" or virtual reference; "image" is denotation or actual reference.
These terms are familiar to us, though it should not be forgotten that it is only with Plato and the Sophists that we see the beginning of the very reflection on language that will later lead to the precise logico-grammatical constructions of the Stoa and the Hellenistic schools. As in book 10 of the Laws or the last part of the Sophist, here in the Seventh Letter Plato presents a theory of linguistic signification in its relation to knowledge. The difficulty naturally begins with the fifth term, which introduces a new element into the theory of signification as we know it. Let us reread the passage: "Each being has three things which are the necessary means by which knowledge of that being is acquired; the knowledge itself is a fourth thing; and as a fifth one must posit the thing itself, which is knowable and truly is." By "fifth" it seems that we should understand the same being with which the excursus begins in saying that "each being has three things." The thing itself would then simply be the thing that is the object of knowledge, and we would thus have found proof for the interpretation of Platonism (which appeared as early as Aristotle) that sees the Idea as a kind of useless duplicate of the thing. Moreover, the list then appears as circular, since what is listed as fifth is what is in truth the first to be named, as the very presupposition from which the whole excursus follows.
Perhaps here we can be aided by philological attention to details, in which, as it has been said, the good God likes to hide himself. At this point the Greek text to be found in modern editions (in Burnet's version, which was in some respects exemplary for all following editions, but also in Souilhé's more recent text) reads: pempton d'auto tithenai dei ho dē gnōston te kai alēthes estin, "and as a fifth one must posit the thing itself, which is knowable and truly is." But the two principal codices on which both scholars base their editions, that is, the Parisinus graecus of 1807 and the Vaticanusgraecus 1, contain a slightly different text, which instead of dei ho ("one must . . . which") has di'ho ("by which"). If we restore the text of the codices by writing di'ho, the translation becomes, "[one must] posit the fifth, by which [each being] is knowable and truly is." 2
In the margin of this text, a twelfth-century hand had noted dei ho as an emendation, and modern editors based their text on this variant. But the codex that Marsilio Ficino had before him for his Latin translation of the works of Plato still respected the text of di'ho, for Ficino's translation reads as follows: quintum vero oportet ipsum ponere quo quid est cognoscibile, id est quod agnosci potest, atque vere existit.
What then changes, what is the significance of this restoration of the original text? Essentially that the thing itself is no longer simply the being in its obscurity, as an object presupposed by language and the epistemological process; rather, it is auto di'ho gnōston estin, that by which the object is known, its own knowability and truth. Even if it is inexact, the marginal variant followed by modern editors is not erroneous. The scribe who introduced it (and we have reason to think it was not an inexpert scribe) was most likely concerned with the risk that knowability itselfthe Idea--would be, in turn, presupposed and substantialized as another thing, as a duplicate of the thing before or beyond the thing. The thing itself--hence the term auto as the technical designation of the Idea--is not another thing but the thing itself, not, however, as supposed by the name and the logos, as an obscure real presupposition (a hypokeimenon), but rather in the very medium of its knowability, in the pure light of its self-manifestation and announcement to consciousness.
The "weakness" of logos therefore consists precisely in the fact that it is not capable of bringing this very knowability and sameness to expression; it must transform the knowability of beings that is at issue in it into a presupposition (as a hypo-thesis in the etymological sense of the word, as that which is placed beneath).
This is the sense of the distinction between on and poion, between Being and its qualification, which Plato insists on several times in the epistle (342 e 3; 343 b 8-c 1). Language--our language--is necessarily presuppositional and objectifying, in the sense that in taking place it necessarily decomposes the thing itself, which is announced in it and in it alone, into a being about which one speaks and a poion, a quality and a determination that one says of it. Language sup-poses and hides what it brings to light, in the very act in which it brings it to light. According to the definition contained in Aristotle (which is also implicit both in Sophist, 262 e 6-7, and in the modern distinction between sense and reference), language is thus always legein ti kata tinos, saying something-onsomething; it is therefore always pre-sup-positional and objectifying language. Presupposition is the form of linguistic signification: speaking kat' hypokeimenou, speaking about a subject.
The warning that Plato entrusts to the Idea is therefore that sayability itself remains unsaid in what is said and in that about which something is said, that knowability itself is lost in what is known and in that about which something is known.
The specific problem that is at issue in the letter, and that is necessarily the problem of every human discourse that wants to make a subject out of what is not a subject, is therefore: how is it possible to speak without sup-posing, without hypo-thesizing and subjectifying that about which one speaks? How is it thus possible legein kat'auto, to speak not by means of a presupposition but absolutely? And since the field of names is, for the Greeks, that which is essentially said kat'auto, can language give reasons (logon didonai) for what it names, can it say what the name has named?
Even the earliest commentators understood that something like a con tradiction is implicit in this problem. We possess a gloss of a late Platonic scholiast that says more or less the following: "Why is it that in the Phaedrus the master gives little value to writing and yet, in having written, in some way holds his own work to be valuable? In this too," the scholiast says, "he wanted to follow the truth. Just as the divinity wanted to create both invisible things and things that fall under our gaze, so he also wanted to leave some things unwritten and others things written." This question certainly holds for the Seventh Letter as well, in which Plato, writing of what concerns him most and what cannot be written about, seems to challenge the weakness of the logos and in a sense to betray himself. And it is certainly not a vain jest that, in another letter, he ends by rejecting the authorship of the dialogues circulating under his name, stating that they are the work of "a Socrates become fair and young." 3 Here the paradox of Plato's written works momentarily flashes up before us: in a letter that the moderns have often taken to be apocryphal, he declares his dialogues to be inauthentic, attributing them to an impossible author, Socrates, who is dead and has been buried for many years. The character about which the text speaks now takes the place of the author in the dialogues in which he appears. The earliest and sharpest critics, such as Demetrius and Dionysius, observe that Plato's style, which is limpid in the earlier dialogues, becomes darker, swollen (zofos) and paratactic (eperriptai allēlois ta kōla aph' eterō heteron, "the phrases are hurled one upon the other," Demetrius writes) when he confronts the subjects dearest to him.
By a curious coincidence, the weakness of language that is called into question by the father of Western metaphysics seems to prophesy from a distance of two thousand years the difficulty implicit in the metaphysical character of our language, which so burdens the writing of the late Heidegger. But in Plato the weakness of the logos does not found a mystical status of the Idea; on the contrary, it renders possible the coming to speech of speech, for the sake of helping speech (logōi boēthein), which in the Phaedrus (278 c 6) is described as the authentic task of philosophical presentation. Here the risk is that the nonthematizability that is at issue in the thing itself will be in turn thematized and presupposed once again in the form of a legein ti kata tinos, a speaking about that about which it is not possible to speak. The thing itself is not a simple hypostasis of the name, something ineffable that must remain unsaid and hence sheltered, as a name, in the language of men. Such a conception, which is implic itly refuted at the end of the Theatetus, still necessarily hypothesizes and sup-poses the thing itself. The thing itself is not a quid that might be sought as an extreme hypothesis beyond all hypotheses, as a final and absolute subject beyond all subjects, horribly or beautifully unreachable in its obscurity. We can, in truth, conceive of such a nonlinguistic thing only in language, through the idea of a language without relation to things. It is a chimera in the Spinozian sense of the term, that is, a purely verbal being. The thing itself is not a thing; it is the very sayability, the very openness at issue in language, which, in language, we always presuppose and forget, perhaps because it is at bottom its own oblivion and abandonment. In the words of the Phaedo (76 d 8), it is what we are always disclosing in speaking, what we are always saying and communicating, and that of which we nevertheless are always losing sight. The presuppositional structure of language is the very structure of tradition; we presuppose, pass on, and thereby--according to the double sense of the word traditio--betray the thing itself in language, so that language may speak about something (kata tinos). The effacement of the thing itself is the sole foundation on which it is possible for something like a tradition to be constituted.
The task of philosophical presentation is to come with speech to help speech, so that, in speech, speech itself does not remain presupposed but instead comes to speech. At this point, the presuppositional power of language touches its limit and its end; language says presuppositions as presuppositions and, in this way, reaches the unpresupposable and unpresupposed principle (arkhē anypothetos) that, as such, constitutes authentic human community and communication. As Plato writes in a decisive passage of a dialogue that presents more than mere affinities with the "extravagant myth" of the Seventh Letter:
Understand then that by the other section of the intelligible I mean what language itself [auto ho logos] touches by the power of dialogue, hypothesizing not by principles [archai] but truly by hypotheses, underpinnings, footings, and springboards, so that it reaches the principle of all things, touching it, and, once again holding to the things near it, returns toward the end, being concerned not with the sensible, but with the Ideas, through the Ideas, toward the Ideas, so that it may end with the Ideas. 4
I realize that I may have gone beyond the task that I set myself; I may be guilty, in some way, of precisely the human folly against which the myth of the Seventh Letter warns us (344 d 1-2): the folly of carelessly consigning one's own thoughts about the thing itself to writing. It is therefore appropriate that I end here, to turn more cautiously to the preliminary historiographical matter that I raised earlier.
We have seen that the digression of the Seventh Letter contains a treatment of the Idea in its relation to language. The determination of the thing itself is, indeed, carried out in close relation with a theory of linguistic signification, one that may constitute the first organic exposition of the material, if in an extremely abbreviated form. If this is true, we should then be able to follow its traces in the Greek reflection on language that immediately follows it. One instantly thinks of the text that, for centuries, determined all reflection on language in the ancient world, Aristotle De interpretatione. Here Aristotle presents the process of linguistic signification in a way apparently without relation to the Platonic digression. "What is in the voice [ta en tēi phōnēi]," he writes,
is the sign of affections in the soul [en tēi psychēi]; what is written [ta graphomena] is the sign of what is in the voice. And just as letters are not the same for all men, so it is with voices. But that of which they are signs, that is, affections in the soul, are the same for all; and the things [pragmata] of which the affections are semblances [homoiōmata] are also the same for all men. 5
A more attentive examination, however, shows precise correspondences with the text of the Platonic excursus. The tripartite division by which Aristotle articulates the movement of signification (en tēi phōnēi, en tēi psychēi, pragmata) textually recalls the Platonic distinction between what is en phōnais (name and logos), what is en psychais (knowledge and opinion) and what is en sōmatōn skhēmasin (sensible object) (Epistle VII, 342 c 6). In view of these affinities with the Platonic epistle, the disappearance of the thing itself in De interpretatione is all the more noticeable. In Aristotle, the thing itself is expelled from hermēneia, the linguistic process of signification. When, later, it momentarily returns in the philosophy of language (as in Stoic logic), it will be so estranged from the original Platonic intention as to be practically unrecognizable.
Aristotle's hermēneia is therefore defined in opposition to the Platonic list, of which it constitutes both a repetition and a refutation. The decisive proof of this polemical distinction is precisely the appearance in the Aristotelian text of grammata, letters. Even ancient commentators wondered about the apparently incongruous appearance of a fourth inter preter alongside the other three (voices, concepts, things). If one keeps in mind that the Platonic excursus aimed to show precisely the impossibility of writing the thing itself and generally the unreliability, for thought, of every written discourse, the marked difference between the two texts is even more evident.
Expelling the thing itself from his theory of signification, Aristotle absolves writing of its weakness. In the place of the thing itself, in the Categories there appears protē ousia, first substance, which Aristotle defines as that which is said neither about a subject (kat' hypokeimenou, by means of a presupposition) nor in a subject. What does this definition mean? First substance is not said on the basis of a presupposition; it does not have presuppositions, because it is itself the absolute presupposition on which all discourse and knowledge are founded. It alone--as name--can be said kat' auto, by itself; it alone--not being in a subject--clearly shows itself. But in itself, as individuum, it is ineffable (individuum ineffabile, according to the formulation of medieval Aristotelianism) and cannot enter into the linguistic signification that it founds, except by abandoning its status as deixis and becoming universal predication. The "what," ti, that was at issue in the name is subsumed into discourse as a kata tinos, "that about which" something is said. They--both the what and the about which--are therefore the same thing, which can be grasped as to ti ēn einai, the Being-the-what-that-was. In this logico-temporal process, the Platonic thing itself is removed and conserved or, rather, conserved only in being removed: e-liminated.
This is why the gramma appears in De interpretatione. An attentive examination shows that in the hermeneutic circle of De interpretatione, the letter, as the interpreter of the voice, does not itself need any other interpreter. It is the final interpreter, beyond which no hermēneia is possible: the limit of all interpretation. This is why ancient grammarians, in analyzing De interpretatione, said that the letter, which is the sign of the voice, is also stoikheion tēs phōnēs, that is, its element. Insofar as it is the element of that of which it is a sign, it has the privileged status of being an index sui, self-demonstration; like protē ousia, of which it constitutes the linguistic cipher, it shows itself, but only insofar as it was in the voice, that is, insofar as it always already belongs to the past.
The gramma is thus the form of presupposition itself and nothing else. As such, it occupies a central place in all mysticism, and as such, it also has a decisive relevance in our time, which is much more Aristotelian and mystical than is usually believed. In this sense--and only in this sense-Aristotle, and not Plato, is the founder of Western mysticism, and this is why Neoplatonism could formulate the accord between Plato and Aristotle that lay at the basis of its school.
Insofar as language bears within it the ontological structure of presupposition, thought can immediately become writing, without having to reckon with the thing itself and without betraying its own presupposition. Indeed, the philosopher is the scribe of thought and, through thought, of the thing and Being. The late Byzantine lexicon that goes under the name of Suda contains, under the entry "Aristotle," the following definition: Aristotelēs tes physeōs grammateus ēn ton kalamon apobrekhōn eis noun, "Aristotle was the scribe of nature who dipped his pen in thought."
Many centuries later, Hölderlin unexpectedly cited this phrase from Suda at a decisive point in his annotations (Anmerkungen) to his translation of Sophocles, namely, in his attempt to explain the sense and nature of Darstellung, tragic presentation. The citation, however, contains an amendment, which Hölderlinian philology, despite its diligence, has not been able to explain. Hölderlin writes: tēs physeōs grammateus ēn ton kalamon apobrekhōn eunoun (instead of eis noun): "he was the scribe of nature who dipped his benevolent pen." Here there is no more dipping of the pen in thought; the pen--that simple material instrument of human writing--is alone, armed solely with its benevolence in the face of its task. To restore the thing itself to its place in language and, at the same time, to restore the difficulty of writing, the place of writing in the poetic task of composition: this is the task of the coming philosophy.

Notes:
1.
Plato, Epistle VII, 340 b 3-7, in Plato, with an English Translation, vol. 7: Timaeus, Critias, Celitopbon, Menexenus, Epistles, trans. R. G. Bury ( Cambridge, Mass.: Harvard University Press, 1952), p. 527. All page citations included in the text of this chapter refer to this edition. Some of the translations have been modified.

2.
Among modern scholars, only Andreae restored the text to its earlier form; see his study on the Platonic Letters in Philologus 78 ( 1923): 34ff.

3.
Plato, Epistle II, 314 c 3-4, in Plato, with an English Translation, 7: 417.

4.
Republic, 511 b 3-c 2, in Plato, The Republic, trans. Paul Shorey ( Cambridge, Mass.: Harvard University Press, 1956), 2:113-15.

5.
Aristotle, De interpretatione, 16 a 3-7. The Greek text is in Aristotle in Twenty-Three Volumes, vol. 1: The Categories, On Interpretation, and Prior Analytics, trans. Hugh Tredennick ( Cambridge, Mass.: Harvard University Press, 1983), p. 114.