by Umberto Eco
True, the Abelard of Sic et non invites us to beware of words used in an unusual sense, of the corrupt state of a text as a sign of a work’s inauthenticity, but the practice will remain imprecise, at least down to Petrarch and the proto-humanists.8
5.2.3. Authentication at the Level of Content
In this case we must decide whether the categories, the taxonomies, the styles of argument, the iconographic configurations, and similar phenomena can be traced back to the cultural universe to which the document is attributed. Even for the modern period such criteria are highly conjectural in nature, though they appeal to relatively accepted notions with regard to the “worldview” typical of a given historical period.
Medieval intellectuals had some idea of content screening; at least during the Scholastic period, they attempted to verify whether a text attributed to a certain author displayed modes of thought in keeping with the cultural universe to which the author belonged. Abelard advises his readers to beware of passages in which the author cites only other people’s opinions, often contradictory, in which the words have a different meaning depending on the author cited. Like Augustine in the De doctrina christiana, Abelard recommends checking the context. But this contextual principle is invalidated by his next recommendation: to give greater weight, in doubtful cases, to the most qualified authority.
Thomas Aquinas takes up the criterion of textual and historical contextuality, giving precedence to usage over the lexicographical meaning; and implicit in this criterion is that the usage be that of the period referred to (Summa Theologiae I, 29, 2 ad 1). Thomas concentrates on the modus loquendi, that is, on the philosophical style, and he is able to establish that at certain points Dionysius the Areopagite or Augustine speak in a certain way because they are following the usage of the Platonists. He goes in search of the intentio auctoris, but his examination is not historical but theoretical. He does not always ask himself whether, at the time of the supposed production of the text, people thought in that way, but rather whether that way of thinking was “correct,” and therefore to be attributed to the supposed doctrinal authority. “In quantum sacra doctrina utitur philosophicis documentis, non recipit ea propter auctoritatem dicentium sed propter rationem dictorum” (“Inasmuch as sacred doctrine makes use of the teachings of philosophy for their own sake, it does not accept them on account of the authority of those who taught them, but on account of the reasonableness of the doctrine”) (In Boet. De Trinitate 2, 3 ad 8).9
Credit is denied to the name of the presumed author (a previous false identification is called into question), but this is done by demonstrating that the alleged author could not have thought what the text says, or think it in the way the text says it.
Let us see how Thomas proceeds in reattributing the De causis, an operation that, when we take the period into account, may be defined as philological—but only in a metaphorical sense. Thomas’s argument goes as follows: until yesterday this book was thought to be by Aristotle, but now we have William of Moerbeke’s translation of the Elementatio theologica of Proclus. Given the similarity of the two texts, we believe that the second is derived from the first, of which it is an Arabic variant, since it comes to us from the Arabic, and its content is not Aristotelian but Platonic. There can be no doubt that we are dealing with an attitude that is already mature, but in this connection it must be observed that these so-called procedures of authentication are based on a concept of authenticity different from our modern criteria.
Thomas repeatedly uses the term authenticus, but for him (and for the Middle Ages in general) the term signifies, not “original,” but “true.” Authenticus expresses its value, its authority, its credibility—not the genuineness of a text’s provenance. Apropos of the De causis he says: “ideo in hac materia non est authenticus” (II Sent. 18, 2, 2, ad 2), but he means that the text is not authentic because it is not in the spirit of Aristotle. In De ver. 1, 1 ad 1, rejecting the attribution of the Liber de spiritu et anima to Augustine, he declares “non est authenticus nec creditur esse Augustini,” but the reasons he gives are purely theoretical (see Chenu 1950: 111).
As Thurot (1869: 103–104) remarks, when explaining texts, the glossators do not attempt to grasp the thought of their author, but to teach the same science that is supposed to be explained therein: “An authentic author, as he was called at the time, cannot be deceived or contradict himself, and neither can he follow a defective plan or be in disagreement with another authentic author.”
5.2.4. Authentication with Reference to Known Fact
In such cases our modern philological disciplines establish whether what the document refers to was indeed the case (or could be known) at the time it was supposedly produced. For example, analyses of the alleged correspondence between Churchill and Mussolini demonstrate the patent falsity of certain letters dated 1945—in spite of the fact that the paper (the material support) is authentic—on the basis of obvious contradictions of known fact. One letter is alleged to have been written from an address where Churchill had not been living for years, another is dated May 7, though in it Churchill refers to events that did not occur until May 10 of that year.
This criterion seems not only “scientific,” but also intuitively obvious. In reality, however, it is very modern. In fact, not only does it presume historical knowledge and the ability to establish on the basis of incontrovertible documentation whether something happened or not in that particular way; it also presupposes that we do not lend credence to the prophetic gifts of the ancient authors.
There is no need to go looking for violations of this principle in the Middle Ages—for the simple reason that we can find a mind-boggling example in the Renaissance. At the height of Humanism, the writings of the supposed Hermes Trismegistus show up at the court of Cosimo de’ Medici, and everyone from Pico della Mirandola to Ficino and beyond is inclined to consider them a product of the ancient world and divinely inspired. The reasoning of these authors, who nevertheless knew both Greek and Hebrew, is not fundamentally different from that of their medieval predecessors: the hermetic texts are divinely inspired because, although they were written before Jesus Christ, they contain the same teachings! They are considered authentically ancient only because they anticipate “prophetically” events (or ideas) that happened later. As we have seen, it will be a good century before Casaubon will turn this criterion on its head: in addition to analyzing expressive forms and forms of content, and demonstrating that the texts of the Corpus Dionysianum contain stylistic traits typical of the Hellenistic period, he will recognize that, if these texts contain echoes of Christian concepts, they must have been composed in the early centuries of the Christian era.
5.3. Three Categories of False Identification
At this point we are in a position to identify three chief forms of false identification.
5.3.1. Strong False Identification
It is asserted (in good or bad faith) that an object Ob is identical (or coincides with) an object Oa, already well-known and famous, where B is an anonymous author, whereas A is an author who is well-known and famous. Oa is instead physically different from Ob and between the two objects there exists merely a relationship of apparent formal homology.
FIRST CASE. A person knows full well that Ob cannot be identified with Oa, because it was produced subsequently by imitation, but still considers the two objects to be equivalent as far as their value and function is concerned and, since he does not possess the notion of authorial originality, he presents the one as identical to the other. This is the case with ingenuous nonfetishistic collecting, as occurred with the Roman patricians who considered themselves aesthetically satisfied with a copy of a Greek statue and were not above labeling it or having it signed “Phidias” or “Praxiteles.” It is the case with the tourists in Florence who admire the David of Michelangelo outside the Palazzo Vecchio, unconcerned that it is a copy of the original preserved elsewhere. A paradoxical variant of this possibility is the authorial fake: the same author A, after
producing Oa, produces, following the same specifications, a perfect double Ob, morphologically indistinguishable from Oa. From the ontological point of view, the two objects are physically and historically distinct, but from the point of view of their aesthetic value they are both equally valuable. Cases of this kind (see the controversy over the fake De Chiricos that some critics believe were painted by De Chirico himself) offer embarrassing food for thought for a critique of the fetishistic concept of the work of art as unicum.
SECOND CASE. A person is aware that Ob is simply an imitation of Oa and cannot be identified with it and does not believe the two objects to be equivalent. But, in bad faith, he pretends (and declares) that Ob is identical to Oa. This is a case of falsification in the strict sense, of a copy identified with the original, or of counterfeiting of currency. The practice has been widespread since classical antiquity, and during the Renaissance collectors commissioned fake coins and statues, often simply for the pleasure of completing their collection.
THIRD CASE. We have a variant of the two previous cases when B transforms Oa into Ob. For example, during the last century the bibliophile Guglielmo Libri manipulated original manuscripts stolen from libraries public and private, dismembering them, altering the notes of provenance and possession, adding false signatures. In a similar way people performed unfaithful restorations on paintings and statues that denatured the work, or they eliminated or covered over parts of the body subject to censure, or broke up the panels of a polyptych. All these operations may have been done in good or bad faith (believing or not believing that Ob was still identical to Oa), or believing or not believing that the work was manipulated in a spirit faithful to the intentio auctoris. In reality, the objects we consider ancient, original, and authentic works of art have instead been transformed by the action of time and by man—and they have undergone amputations, restorations, alterations, loss of color. To this category belongs the neoclassical dream of Greek art as “white,” whereas the original temples and statues were polychrome. In this way, a typology of falsification may lead us to reflect critically on our own ideology of authenticity.
FOURTH CASE. A person is unaware that the two objects are not identical, or believes that Oa and Ob are the same object. Obviously he is not concerned with the problem of their interchangeability and presents Ob as authentic. This was a common state of affairs in the Middle Ages, but it can also occur today in the case of an erroneous authentication made in good faith.
5.3.2. Weak False Identification or Presumption of Interchangeability
Oa and Ob are known to be physically different, but it is agreed that, when described in a certain way and for certain practical purposes, the one is equivalent to the other, and they are presented as completely interchangeable.
This was the case in general in the Middle Ages for all translations. The translation was the only text that supplied information about the original, and it was considered a substitute for the original, even though it was known to be a version from another language (usually unknown). This was also the case for transcription from one codex to another. From the point of view of modern philology these translations and transcriptions were all unfaithful, in addition to which translator and transcriber would consciously alter the text, amputating it or censuring it. To this category we may also assign the various kinds of hidden censure that translations and copies were subject to, and even certain cases of aberrant decoding produced by an annotation that led the copyist to interpret one expression as if it was the same as another.
The Middle Ages was very flexible in its attitude toward translations. In paragraph IV, v, 7, 134, of the De divinis nominibus of Pseudo-Dionysius, Hilduin’s first version translated kalon as bonum and kallos as bonitas. Eriugena translates the first term as bonum but the second as pulchrum; and lastly John the Saracen renders both with pulchritudo and pulchrum. These are substantial differences that reveal, as De Bruyne (1946: 1:5, 2) points out, a profound cultural transformation. But the Saracen himself, in a letter to John of Salisbury (PL 193, 2599), will claim that he translated according to the meaning, not according to the letter. The Saracen was lexically correct, but probably for the wrong reasons, at least in terms of the official lexicography of his day, since, in the following century, Albertus Magnus will continue to debate the two terms and to assert that kalos with one “l” means goodness, not beauty.10
5.3.3. Pseudo-Identification
This is the case of apocryphal or pseudoepigraphical objects. It is asserted that an object Ob is identical to (or coincides with) an object Oa, except for the fact that Oa no longer exists, or never existed, and in any case has never been seen by anyone. Oa is qualified as exceptional, either because of the name of its author or because in reality the tradition has handed down inaccurate information about its supposed existence. To lend credence to a pseudo-identification we have to be somehow familiar with a set of objects a (Oa1, Oa2, Oa3, etc.), all produced by a well-known and famous author A. From set a an abstract type is extracted which does not take into consideration the features of objects a but instead the supposed specifications according to which they were formed, or the way in which A apparently produced them (style, type of materials used, etc.). Ob was produced according to these specifications, and it therefore is asserted that Ob is a previously unknown product by A.
FIRST CASE. Someone is aware that Oa does not exist and is familiar only with Ob. He therefore knows that they cannot be identified with each other. But he believes in good faith that Ob may serve all the purposes that Oa would have served, and as such he presents it in place of Oa, whereas Ob is merely an ersatz of Oa. This is the typical case of the diplomatic forgery (reine formale Falschung). While the historical forgery (reine Falschung) concerns a formally genuine document that contains inexact or invented information (such as the authentic confirmation of false privileges), the diplomatic forgery is a document expressly created to assert privileges that may in fact have really been conceded but whose original documentation has been lost. Examples are the false documents produced by monks to backdate or extend the possessions of their abbeys, where we may suppose that the monks, on the basis of tradition, were convinced that they had truly obtained the privileges in question and were simply attempting to affirm them in a public manner.
SECOND CASE. Someone knows that Oa does not exist, and does not believe that Ob is equivalent and interchangeable with Oa. Nevertheless, in bad faith, he insists on declaring the two objects (one real, the other virtual) to be identical, or on the authenticity of Ob, with intent to deceive. This is the case of the modern diplomatic forgery, of fake genealogical trees produced to confirm otherwise unattested pedigrees, of apocryphal documents produced with malicious intent. This is probably the case of the poem De vetula, produced in the thirteenth century, but immediately attributed to Ovid. We may suppose that the person or persons who placed the Corpus Dionysianum into circulation in the eighth century, attributing it to a disciple of Saint Paul, were instead aware that the work had been fabricated much later, but they nonetheless decided to attribute it to an uncontestable auctoritas. To this category there also belong the cases of attribution to an author by no means well-known and famous, but who becomes so when he is presented as ancient and when characteristics are attributed to him that make him an authority. This is the case with a number of nonexistent chroniclers to whom the Abbot Trithemius attributed spurious works.11
In all of these cases, in addition to the documentary forgery, a historical falsification is also committed, in other words, lies are circulated regarding events of the past. The pseudo-identification is invoked to subrogate the historical lie.
THIRD CASE. Someone is unaware that Oa does not exist and does not know that it is not identifiable with Ob. Therefore that someone has no problem considering them identical. Independently of whether or not he believes in the interchangeability of the two objects, he claims in any case that they are identical, thereby affirming the authenticity of Ob. This appears to have been t
he case with those who thought the Corpus Dionysianum was the work of a disciple of Saint Paul, unaware that it had been produced at a later date, and those who considered the De causis to be a work by Aristotle and not by an Arabic follower of Proclus. It is certainly the case with all those who believed and continue to believe in the authenticity of the book of Enoch, and to the men of the Renaissance who attributed the Corpus Hermeticum not to Hellenistic authors but to a mythical Hermes Trismegistus who supposedly lived before Plato at the time of the Egyptians and could probably be identified with Moses. In the modern period, we have the case of Heidegger (1915) who writes a commentary on a Grammatica Speculativa believing it to be the work of Duns Scotus, while a few years later it will be proven to be the work of Thomas of Erfurt. It goes without saying that a false attribution of this kind also leads to aberrant decoding.
A variant of this case of pseudo-identification is attribution to a pseudo-author: we have only one text Ob, whose author is unknown, and it is decided to attribute it to an author A, information about whom is uncertain. This seems to have been the case with the attribution of the treatise On the Sublime to a certain Pseudo-Longinus.
5.4. What Do We Mean by “Knowing That”?
In sketching this semiotics of falsification we have implicitly made use of an epistemic operator like “knows that” which poses a number of problems. What does it mean to say that someone knows that Oa and Ob are not identical? The only case of false attribution in which we can know that Oa and Ob are not identical is the one in which someone presents us, for example, with a perfect reproduction Ob of the Mona Lisa, when we are standing in front of the original Oa exhibited in the Louvre, and affirms that the two objects are indiscernibly the same object. This is of course an improbable event, but even if was to occur the doubt would remain whether Ob was the authentic Mona Lisa and Oa a fake maliciously (or erroneously) hung on the gallery wall. And what does it mean to know that Oa never existed? Except for the case in which there are irrefutable proofs that Oa once existed and has been destroyed (as is probably true of the Hanging Gardens of Babylon or the temple of Diana at Ephesus), usually the assertion “Oa does not exist” is understood simply to mean “there are no proofs of its existence.”