From the Tree to the Labyrinth

5

Fakes and Forgeries in the Middle Ages

The modern reader, nurtured on philology, is aware that many forgeries were perpetrated in the course of the Middle Ages. But were the people of the Middle Ages similarly aware? Did they recognize the notion of forgery? And if they recognized the notion, was it the same as our own?

In formulating these questions, we find ourselves compelled to analyze a series of terms—like falsification, fake, forgery, false attribution, diplomatic forgery, alteration, counterfeit, facsimile, and so on—that we nowadays take for granted. If we are to decide whether similar concepts existed in the Middle Ages, we are inevitably obliged to take a closer look at our own contemporary concepts.

It is no accident that dictionaries and encyclopedias, in defining falsification, place the emphasis on malicious intent, introducing—without defining them—concepts such as counterfeit, spurious, apocryphal, pseudo, and so on. Webster’s Unabridged Dictionary, for instance, defines forgery as “the act of forging, fabricating or producing falsely; especially the crime of fraudulently making, counterfeiting, or altering any writing, record, instrument, register, note and the like to deceive, mislead or defraud; as the forgery of a document or of a signature.”¹

The dictionaries are also vague on the distinction between spurious, apocryphal, and pseudo. Spurious is used for nonauthentic or falsified works and documents, but also for an illegitimate child born from an adulterous relationship. In the natural sciences, it refers to organs that resemble other organs without having their function. For example, the spurious ribs are two lower ribs on either side of the skeleton that do not reach as far as the sternum; in zoology, the spurious or bastard wing (or alula) is a tuft of accessory flight feathers growing on the first digit of the bird’s wing, behind the wing’s angle, in some cases substituted by a nail or spur; in botany, it indicates an apparatus or organ that resembles another organ with a different structure or function.

In German the same phenomenon is rendered with the prefix pseudo. Webster gives apocryphal as a synonym of spurious (“Apocryphal: various writings falsely attributed … of doubtful authorship or authenticity … spurious”). In fact, apokryphos originally meant occult and secret; apocryphal gospels and other biblical writings got the name because people weren’t allowed to read them—and as such they were excluded from among the canonical books. Hence, “apocryphal” came to signify “excluded from the canon.” Subsequently, late Jewish authors attribute their writings to the ancient prophets, and these books are termed pseudonymous or pseudoepigraphical. It should be observed, however, that Catholics describe the noncanonical books as apocryphal, while the books accepted in the Greek version of the Septuagint are said to be deuterocanonical. For Protestants on the other hand it is the deuterocanonical books that are apocryphal while the ones Catholics call apocryphal are pseudoepigraphical.²

5.1. The Semiotics of Forgery

Given the complexity of the notion of forgery, if we are to understand what might have been considered a forgery in the Middle Ages, we must proceed to clarify the various related concepts.³

5.1.1.

Doubles

The first thing we must consider is the semiosic concept known as replicability. The most complete instance of replicability is the double, a physical token that has all the characteristics of another physical token, at least from a practical point of view, insofar as both possess all the pertinent traits prescribed by an abstract type. In this sense, two chairs of the same model or two sheets of office paper are both doubles of one another, and the perfect homology between the two tokens is established with reference to their type. Doubles do not lend themselves to the deceit of falsification in that every token has the same practical value as every other, and each one can substitute for the other. A double is not identical with another double (in the Leibnizian sense of indiscernibility), in other words, two tokens of the same type are—and are recognized as—two different physical objects. Nevertheless, they are considered interchangeable.

Theoretically speaking, we have two reciprocal doubles when, given two objects, Oa and Ob, their matter displays the same physical characteristics, in the sense of their molecular composition, and their form is similar, in the mathematical sense of congruence (the features to be compared for similarity are determined by the type). But who is to determine the criteria for similarity? The problem of doubles is ontological in theory, but pragmatic in practice. It is the user who decides under which description—that is, from what practical standpoint—the two matters and the two forms are, ceteris paribus, “objectively” similar, and therefore, from the practical point of view, interchangeable. Under a microscopic analysis, or in the light of other chemical tests, it could be proven that two sheets of office paper of different brands display fairly relevant differences, but a normal user habitually sees them as doubles (and hence interchangeable) in every respect.

5.1.2. Pseudo-Doubles

We have a case of pseudo-doubles when only one token of the type (the privileged token) takes on a special value in the eye of one or more users, for one or all of the following reasons: (i) on account of temporal priority, such as occurs for instance when the first product off the assembly line of particular model of automobile (if it can be identified as the first) is displayed in a museum as a unique specimen;⁴ (ii) because that particular token contains evidence of previous possession, as occurs in the case of a copy of a book with an inscription by the author or the signature of an illustrious former owner; (iii) because that token has been used in a special context (this would be the case with the Holy Grail, the chalice used by Jesus Christ at the Last Supper, if it could ever be discovered and authenticated); (iv) because the particular token is of such material and formal complexity that no attempt to imitate it can reproduce all the characteristics recognized as relevant (a typical case would be an oil painting on canvas painted in a particular style with special paints, so that the chromatic shadings, the microscopic grain of the canvas, the flow of the brushstrokes—all features judged indispensable to the total fruition of the object—can never be completely imitated.⁵ In all of the above cases, for various reasons, these “unique” objects become the type of themselves, and any reproduction of these objects, when not honestly presented as a facsimile or imperfect copy produced for a didactic or documentary purpose, is made with a false identification in mind.⁶

5.1.3. False Identification

We have false identification when, given a hypothetical object Oa, produced by author A in historical circumstances t₁, and, given another object Ob, produced by author B in historical circumstances t₂, somebody (an individual or a group) decides that Ob is identical with Oa, to the point of being indiscernible. In the concept of falsification the malicious intentions of the falsifier are generally implicit. The problem of malice on the part of B, the author of Ob, seems to us irrelevant: he is fully aware that Ob is not identical to Oa, but he may have produced it with no intention to deceive, as an exercise, as a joke, or by mere chance. The Constitutum Constantini (Donation of Constantine) was probably first produced as a rhetorical exercise, and it was only in later centuries that (in good or bad faith) it came to be considered authentic (see De Leo 1974). What interests us more is the intention of the person performing the false identification (the Identifier) who asserts that Oa and Ob are identical (of course, in a case of malice aforethought, the Identifier and author B of Ob may be one and the same person).

Historical forgery does not belong in this category. It concerns a document Ob, produced by B, who is entitled to produce it as his own, but whose purpose in producing it is to assert (in a mendacious fashion) something inexact or invented. This is the case, for instance, when someone writes a letter bearing false witness, a report that misrepresents the results of a scientific experiment, a dispatch or communiqué issued by a government that lies about the results of an election (electoral fraud), and so on. A historical forgery is an instance of a deliberate lie and in this sense it is to be distinguished from a diplomatic forgery, which we will come to later.⁷

5.2. Difficulties of Authentication Procedures

In order for a process of false identification to occur a culture must have criteria, considered somehow objective, by which to establish indiscernibility or equivalence between objects, and therefore criteria for establishing the authenticity of an object Ob. These criteria can be valid (i) for objects that were not produced for communicative purposes, such as paleontological finds, objects in use in archaic or primitive cultures (which can be interpreted as signs, symptoms, traces, or clues to events distant in space and time); or (ii) for objects produced for explicitly communicative purposes (documents, visual works of art, hieroglyphic inscriptions, epigraphs, etc.). Both kinds of objects are generally understood to be “documents,” though objects belonging to type (ii) are considered both for their expression and their content, while objects belonging to type (i) are evaluated only for their expression, seeing that the content (or meaning) attributed to them by the addressee did not exist for the sender (the archaic producer of an iron knife blade undoubtedly intended to signify the practical function of the object he was constructing, but only the modern archeologist reads that knife as a sign of the fact that, when it was produced, people knew how to work iron).

The contemporary disciplines of identification (which we will refer to generically as philological disciplines) recognize four methods of authentication. We will see, case by case, that the criteria available to medieval culture were far more vague.

5.2.1. Authentication at the Level of the Material Support of the Text

We have physicochemical methods for determining the period of fabrication and the quality of the material support (parchment, paper, canvas, wood, etc.). Nowadays these methods are considered sufficiently scientific, and therefore intersubjectively verifiable, but the medieval scholar almost never had the opportunity to encounter original documents in their original language (even the translators were working from manuscripts at a considerable remove from their archetypes), and all they knew of past civilizations were seriously contaminated ruins. Christianity discovers history (the sequence creation-original sin-redemption-parousia), but not historiography. It knows the past solely through the information handed down by tradition. The legal opinions handed down in the High Middle Ages ascertaining the counterfeit nature of the documents produced by one of the litigating parties confine themselves at best to a discussion of the authenticity of the seal. Remi of Trèves asks Gerbert d’Aurillac (the future pope Sylvester II) to send him one of his leather armillary spheres, and Gerbert (an enthusiast of the classical authors) asks for a copy of Statius’s Achilleid in exchange. Remi sends it to him; but the Achilleid was left unfinished by its author. Gerbert is unaware of this and accuses Remi of sending him a defective manuscript and, to punish him, sends him an inferior painted wooden sphere. Gerbert had no accredited sources for knowing the physical conditions of the original manuscript (see Havet 1889: 983–997 and Gilson 1952: 228–229).

The cautionary tale of the reception and translations of the Corpus Dionysianum is an episode worthy of reflection. When Byzantine emperor Michael II the Stammerer sent it as a gift to Frankish king Louis the Pious in 827 as the work of a disciple of Saint Paul who was the first bishop of Paris, no one thought to question its authenticity. The testimony of the donor, the prestige of the alleged author, the interest of the text—all were sufficient guarantees. Scotus Eriugena had doubts about the identity of Paul’s disciple and the first bishop of Paris, but not about the venerable age of the text.

5.2.2. Authentication at the Level of Textual Manifestation

The form of the document must be in keeping with the rules of formation of the period to which it is attributed. The first example of philological analysis based on the form of the expression was provided in the fifteenth century by Lorenzo Valla (De falso credita et ementita Constantini donatione declamatio, XIII), when he demonstrated that the use of certain linguistic expressions in Latin was completely implausible at the beginning of the fourth century A.D. Isaac Casaubon (De rebus sacris et ecclesiasticis exercitationes, XIV) proved that the Corpus Hermeticum was not a Greek translation of an ancient Egyptian text, because it did not contain a single trace of Egyptian idiomatic expressions. Modern philologists have shown that the Hermetic Asclepius was not translated, as was once believed, by Marius Victorinus, because in all his writings Victorinus always put the conjunction etenim at the beginning of the sentence, whereas in the Asclepius the word occurs in the second position twenty-one times out of twenty-five.

According to the semiosic system, recourse is made to paleographic, epigraphic, lexicographic, grammatical, iconographic, and stylistic and other criteria. These methods are today judged sufficiently scientific, even when based on conjecture. The Middle Ages had no paleographic criteria, and its lexicographic, grammatical, and stylistic criteria were fairly vague. Men like Augustine and Abelard, and eventually scholars like Thomas Aquinas, recognized the problem of establishing the reliability of a text on the basis of its linguistic features. But, apropos of the text of the Bible, Augustine, who had small Greek and less Hebrew, in the pages where he discusses the technique of emendatio, advises at most to compare the various Latin translations with each other, in order to make a conjecture, taking account of the differences, about the “correct” reading of the text. He is looking for a “good” text, not the original text, and he rejects the idea of checking the Hebrew version because he believes it has been manipulated by the Jews: hence, not only does he not go back to the presumed original, he mistrusts it. Better a translation inspired by God that an original corrupted by a malicious intent (De doctrina christiana 2, 11–14).

As Marrou (1958) remarks, none of his commentaries presupposes a preliminary effort to establish a critical text. There is no analysis of the manuscript tradition. Saint Augustine is content to compare the largest possible number of manuscripts and to take into consideration the largest possible number of variants.

When Saint Jerome’s translation ex hebraeo conflicts with that of the Septuagint, Augustine tends to suspect Jerome’s translation, because he considers the Septuagint divinely inspired. He never chooses the Vulgate over the Septuagint. In the De civitate Dei (15, 10–11), in calculating the age of Methuselah, the text of the Septuagint (but not the Vulgate) is contradictory, since it has Methuselah die after the Flood, but Augustine refrains from committing himself, suggesting the hypothesis of a correction introduced by the perfidious Jews to undermine the confidence of Christians vis-à-vis the Septuagint version. It is curious that Augustine should think that the Hebrew original might be corrupt (a useful suspicion on the part of a philologist), while he is not overly concerned over the corruptness of the translations, convinced that he can resolve the issues with a bland comparative approach, in which the last word will be uttered not by philology but by a righteous will to interpret and fidelity to traditional knowledge (see Marrou 1958: 432–434).

Bede and other authors analyze the rhetorical figures of Holy Scripture, but they are ignorant of the Hebrew original, and the language they are analyzing is that of a translation. It is not until the thirteenth century that an effort will be made to return to the Hebrew original with the help of converted Jews (see Chenu 1950: 117–125 and 206).

In any case, etymological practice has much to teach us about the weakness of medieval philology, whether the etymologies in question be those of Isidore of Seville or Virgil of Toulouse. Medieval etymology has nothing to do with the history of the lexicon. It is philosophical, theological, moral, or poetic. Every medieval etymology is, from the etymological point of view, a fake.

As for their insensitivity to language, the case of the thirteenth-century Modistae (see Chapter 7) is exemplary: all of their speculative grammar is an example of philological highhandedness. They attempt to elaborate a general theory of language on the basis of a single language, Latin. They do not believe that other languages displaying other grammatical (and therefore mental) structures exist. They identify modus essendi and modus significandi. Their ethnocentric impermeability is equal only to that of those twentieth-century Anglo-Saxon linguists who construct theories of linguistic universals on the basic of a single language, English.

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 the 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.”

Modern philology has developed techniques of identification to establish whether an Ob is identical to an Oa, but these procedures presuppose that we know the properties Oa has or should have. Now, the techniques by which we establish the characteristics of Oa are the same as those by which we identify Ob. In other words, in order to say that a reproduction of the Mona Lisa is not authentic, somebody has to have analyzed and authenticated the original Mona Lisa using the same techniques used to decide that the reproduction of it is different. For modern philology the traditional evidence that the Mona Lisa in the Louvre was put there, let’s say, by Leonardo right after painting it is not enough. This fact must be proven by means of documents, but for these documents too the question of their authentication must be posed. And if there is any doubt about the documents, the presumed original of the Mona Lisa is analyzed to decide whether its material and morphological attributes lead us to conclude that it was painted by Leonardo.

Our modern culture, therefore, must assume that (i) a document authenticates traditional information and not vice versa; (ii) authenticity means historical primitivity and authorial originality (this is the only way to establish the priority of Oa over Ob); and (iii) primitivity and originality are established by considering the object as a sign of its origin, and the techniques of authentication described in section 5.2 are applied to this end.

These checks call for scientific and historical knowledge of which the Middle Ages had only a vague and ambiguous grasp, for reasons intimately connected with its concept of historical truth.

5.5. Historical Truth, Tradition, and

Auctoritas

The Middle Ages could not conceive of a document that would authenticate traditional data because the only reliable form of documentation it possessed was traditional data.

The Middle Ages could only argue based on the testimony of the past, and the past had chronological abscissas that were quite vague. The medieval procedure of recourse to authority has the form of a synecdoche: an author or a single text stands for the globality of tradition and always functions outside of any context. Le Goff (1964: 397–402) has remarked that the medieval form of wisdom is folkloristic, and is symbolized by the proverb. Feudal law and practice are sanctioned by custom.

The same Le Goff cites a 1252 lawsuit between the servants of the chapter of Notre Dame de Paris in Orly and the canons: the canons say the servants must pay tithes because tradition requires it. The oldest inhabitant of the region is consulted and he says that it has been that way “a tempore a quo non extat memoria” (“from time immemorial”). Another witness, the archdeacon Jean, affirms that he has seen certain ancient documents in the chapter which attest to the existence of the custom, and the chapter has put its faith in these documents out of respect for the antiquity of the writing. No one of course thought to check the existence, let alone the nature, of the documents: it was sufficient to hear they existed, for centuries.

For the Middle Ages, the problem of tradition, in historiography and hermeneutics, is that it does not have to be reconstructed: it is already given from the beginning; it must simply be recognized and interpreted in the proper way.

Apart from the data of tradition, only one document is recognized, and it is the text (translated) of the Holy Scriptures. Other documents are not distinguished as original and nonoriginal: they have either been handed down or they don’t exist. If they have been handed down, they are true only insofar as they agree or can be made to agree with the truth of Scripture: “Certus enim sum, si quid dico quod Sacrae Scripturae absque dubio contradicat, quia falsum est” (“For I am certain that, if I say anything which clearly opposes Holy Scripture, it is false”) (Anselm, Cur Deus homo, 1, 18, PL 153, 38).

Still, the problem is not so simple, because, in order to establish the truth of Scripture, it must be correctly interpreted. After Origen proposed the principle of the complementarity of the two testaments and their parallel reading, the problem arose of how to legitimate their interpretations. On the one hand a correct interpretation must legitimize the Church, but on the other what decides whether and how an interpretation is correct is the interpretive tradition, legitimized by the Church as the guardian of truth: an embarrassing situation, and the origin of every theory of the hermeneutical circle (see Compagnon 1979).

This is why the Middle Ages must amass a treasury of authoritative opinions, or auctoritates. In the course of the philosophical and theological debate, authority materializes in the form of quotes that become “authentic” opinions and therefore authoritative in themselves. They are clarified, when they are obscure, by their glosses, but these too must come from an “authentic” author.

As Grabmann remarks (1906–1911), when it came to the explanation of Scripture, historical grammatical interpretations or independent research on the concepts and connections of the biblical text carried no weight; what counted were above all collections of passages extrapolated from the Fathers of the Church. Pre-Scholastic theological literature “is placed under the sign of reproduction,” and appeals to florilegia and catenae. But little by little the original manuscripts of the Fathers are neglected or lost, and their opinions survive only in the florilegia. When we consider that this process occurs through free transcriptions and translations, we can see how the modern idea of authenticity could find itself in considerable difficulty.

Furthermore, the florilegia are arranged for the most part in alphabetical order, which excludes the kind of systematic classification that might have made for comparison and discussion of contradictory passages. With the twelfth century, the florilegia and traditional opinions are supplemented by sententiae modernorum magistrorum, even though these so-called modern masters are such only by academic convention (as authors of glossae magistrales), and Thomas often dares to contradict them (“haec glossa magistralis est et parum valet,” [“this is a master’s annotation and has little value”] In I Timeum 5, 2).¹²

To the anarchy of the authorities, the Middle Ages proved incapable of opposing a practice of verification of historical originality. Scrutiny (and the dialectical discussion intended to resolve contradictions) was not philological but philosophical. Hence the decision, asserted without hypocrisy in the twelfth century, to treat authorities with a pinch of salt. “Authority has a nose of wax, in other words, it can be bent in different directions” [“Auctoritas cereum habet nasum, id est in diversum potest flexi sensum,” Alain de Lille, De fide catholica 1, 30]). Authorities must be accepted, but, given their insufficiencies and contradictions, they must be interpreted reverently, exponere reverenter, and, as Chenu notes (1950: 122), we should make no mistake over the meaning of this expression: what we are dealing with are small but efficacious adjustments, fine-tuning, rectifications to the meaning of the text.

5.6. On the Shoulders of Giants

Bernard of Chartres, as we know, supplied the moral and historical justification for these interpretive liberties, with his famous aphorism that compared contemporary thinkers to dwarves standing on the shoulders of giants.¹³ But the same idea (if not the metaphor of the dwarves) appears six centuries earlier in Priscian, and this brings us to the question of whether the aphorism is modest or presumptuous in its intent. In fact it can be interpreted in the sense that what we know today, though we may know it somewhat better, is what the ancients have taught us, or, alternatively, that, however much we owe to the ancients, we know far more than they did. A similar aphorism, that appears in Saint Bernard of Clairvaux (Bernardus Carnotensis) and speaks of gleaners following in the footsteps of the reapers, leaves no room for doubt, because the gleaners gather only the gleanings left behind by the reapers. Where Priscian stood remains ambiguous: for him it seems that the moderns are more perspicacious than the ancients, though not necessarily more learned.¹⁴

But perhaps we should be debating not the meaning of the aphorism but how it has been interpreted in various historical periods. What does William of Conches mean when, commenting on the aphorism, he declares that the moderns are “perspicaciores” (“more perspicacious”) than the ancients? It is no accident that, taking Newton as his point of departure, Merton (1965) sees the aphorism as decisive in the modern debates over influence, collaboration, borrowing, and plagiarism. But the notion of plagiarism, and the idea of staking one’s life on being or not being the first to see something, can exist only in a period in which what is prized in every field of discourse is originality, or in the spirit of that modernity characterized by Maritain with the telling formula to the effect that, after Descartes, every thinker becomes a “debutant in the absolute.” In the Middle Ages that was not how it was at all.

In the Middle Ages what was true was true because it had been upheld by a previous authority, to the point that, if one suspected that the authority had not espoused the new idea, one proceeded to manipulate the evidence, because authority has a nose of wax. It comes naturally to the Middle Ages to employ the aphorism, because the mode of discussion typical of the period is the commentary or the gloss. One must always take a giant as one’s point of departure. But it is up for grabs whether a medieval thinker using the aphorism is vindicating the superiority of the moderns or arguing for the continuity of knowledge.

To read the aphorism in a Hegelian sense we do not have to wait for Hegel, but neither must we assume that Bernard thought like Newton. Newton knew full well that, since Copernicus, a revolution in the universe was under way; Bernard didn’t even know that revolutions in knowledge were possible.

Indeed, since one of the recurrent themes of medieval culture is the progressive senescence of the world, Bernard’s aphorism could be interpreted to mean that, given that mundus senescit (the world is getting older and older), and inexorably at that, the best we can do is to play up some of the advantages of this tragedy.¹⁵

On the other hand, Bernard, following Priscian, uses the aphorism in the context of a debate on grammar, in which what is at stake are the concepts of knowledge and imitation of the style of the ancients. Nothing to do then with notions like the cumulative nature and progress of theological and scientific knowledge. Still, Bernard (our witness is still John of Salisbury) scolded those among his pupils who slavishly imitated the ancients, saying that the problem was not writing like them, but learning from them to write as well as they did, so that, in the future, “someone will be inspired by us as we are inspired by them.”¹⁶ Therefore, though not in the same terms as we read it today, an appeal to independence and courage was nonetheless present in his aphorism. And it is not without significance that John of Salisbury takes up the aphorism no longer in the context of grammar but in a chapter in which he is discussing Aristotle’s De interpretatione.

A few years earlier Adelard of Bath had inveighed against a generation that considered acceptable only the discoveries made by the ancients, and in the coming century Siger of Brabant will declare that auctoritas by itself is not enough, because we are all men exactly like those we are inspired by, and therefore “why should we not devote ourselves to rational research like them?” (Beonio-Brocchieri Fumagalli 1987: 232). We are clearly on the threshold of modernity. But we have a long way to go as far as the concept of originality and the neurosis of plagiarism are concerned.

5.7.

Tamquam ab iniustis possessoribus

The Middle Ages copied without acknowledgment because that was the way it was done and ought to be done. What’s more, a notion akin to that contained in the aphorism was anticipated by Augustine and developed by Roger Bacon, when he said that if we find good ideas in pagan texts we are entitled to appropriate them as ours “tamquam ab iniustis possessoribus,” (“as it were from unjust possessors”) because, if the ideas are true, then Christian culture has every right to them. This explains why medieval notions of forgery and what is fake are very different from our own.

True, the falsification of auctoritates is an act of critical freedom that reaffirms the principle of discovery against every kind of dogmatic constraint. But this liberation is obtained at the expense of what we would define today as “philological correctness.” If the dwarf is to see further than the giant he can and must adjust the giant’s thought to show that innovation does not contradict tradition. Non nova sed nove (“Not new things, but in a new way”). This is why medieval culture could not avoid a casual approach to philology.

Let us close with a significant example. Thomas’s choice of translations seems never to be inspired by philological considerations. His commentary on the De interpretatione follows the translation by Boethius, despite the fact that he already had William of Moerbeke’s new version available to him, and without realizing that Boethius was guilty of a misreading of considerable interpretive importance. In De interpretatione 16a Aristotle says that words are symbola of the passions of the soul, but shortly thereafter he adds that they can also be taken as semeia of the same passions, and hence as symptoms. The passage can be explained as meaning that words are conventional symbols, but they may also be interpreted as symptoms of the fact that the speaker has something on his or her mind. As we already saw in Chapter 4 on the barking of the dog, Boethius translates both Greek terms with nota (a fairly vague multipurpose expression), which leads Thomas to interpret both cases with the word signum—a choice that seriously compromises a correct reading of the text.

But note what happens with Roger Bacon, who was so convinced that, in order to snatch the truth from the infidel, “tamquam ab iniustis possessoribus,” we have to know languages, to be able to check the translations—an ideal shared by Robert Grosseteste and in general by the Oxford Franciscans (“Cum ignorat linguas non est possibile quod aliquid sciat magnificum, propter rationes quam scribo, de linguarum cognitione,” Opus Minus, p. 327). Bacon knows Greek and perhaps he realizes Boethius’s error. But even after realizing it, for reasons that have to do with his own theory of signs, he continues to see the relationship between words and things as purely symptomatic, as if Aristotle had used the term semeion in both instances (De signis, V, 166).

Bacon is aware that a translation Ob is not the equivalent of the original Oa, but he has no qualms in transforming Oa into a third text that simply turns Ob upside down. He certainly acted without any clear intention to deceive, and felt authorized to do what he did because he was convinced that in so doing he was better serving the interests of truth. But the truth was his truth, not the truth of the original text.

It is episodes like this that lead us to conclude that, though there were forgeries in the Middle Ages, what was missing was the awareness of forgery. Medieval notions of true and false attribution and manipulation of a text were not the same as ours.

5.8. Conclusions

We could say, as tradition has it, that the new philological awareness begins with Petrarch, and subsequently with Lorenzo Valla. But the fact that this awareness surfaced does not mean that European culture changed its attitude toward its sources overnight. The proof, furnished by Casaubon, of the Hellenistic origins of the Corpus Hermeticum appeared at the beginning of the seventeenth century, but even afterward, and for a considerable length of time, most of European culture continued to believe in the text’s antiquity.

We would be better advised to reflect on the regeneration of the processes of falsification in the contemporary world. Setting aside the fabricators who continue to repeat the time-honored counterfeiting techniques (false attributions, fake genealogical tables, copies of paintings), we find ourselves faced, in the political universe and in the mass media, with a new form of falsification. Not only do we have false information, but also apocryphal documents, placed in circulation by a secret service or a government or an industrial group, and leaked to the media, in order to create social turmoil, confusion in public opinion. We speak of “false information,” without appealing to epistemological considerations, because the news is bound to be discovered as false sooner or later. Indeed we might say that it is disseminated as true precisely in order for it to be revealed as false a little time later.

Its purpose in fact is not to create a false belief but to undermine established beliefs and convictions. It serves to destabilize, to throw suspicion upon powers and counterpowers alike, to make us distrust our sources, to sow confusion.

We conclude then that the people of the Middle Ages falsified in order to confirm their faith in something (an author, an institution, a current of thought, a theological truth) and to uphold an order, whereas our contemporaries falsify in order to create distrust and disorder. Our philological age can no longer permit itself falsifications that present themselves as truths because it knows they will be unveiled in no time; and it operates instead by spreading falsifications that have no fear of philological examination, because they are destined to be unmasked immediately. We are not dealing with an isolated fake that masks, hides, and confuses, and to that end endeavors to seem “true.” It is the quantity of falsifications recognizable as such that functions as a mask, because it tends to undermine the reliability of all truth.

We do not know how the people of the Middle Ages, with their ingenuous concept of authenticity, would have judged this brash and cynical concept we have of noningenuous falsification. One thing is for sure: no historical period has the right to moralize about any other.

A revised version of “Tipologia della falsificazione” [“A Typology of Forgery”], in Setz (1988), originally given at the Internationaler Kongress der MGH, Munich, September 16–19, 1986. My theoretical (rather than historical) essay, “Falsi e contraffazioni,” was developed on the basis of this publication (see Eco 1990a). [Translator’s note: Relevant also is Eco’s entry (published in English) “Fakes in Arts and Crafts” in Eco 2004c (4:3571–3580).

1. Equally unsatisfactory are the German definitions in the Brockhaus Enziklopädie (1968) (“Zweck vorgenommene Nachbildung, Veränderung oder historisch irrefhrende Gestaltung eines Gegenstandes (hierzu Tafeln), eines Kunstwerkes, eines literar. Denckmals, einer Unterschrift usf.”) or the Meyers Grosses Universal Lexikon (“der Herstellen eines unechten Gegenstandes oder das Verändern eines echten Gegenstandes zur Tauschung im Rechtverkehr—dadegen Imitation”). The following definitions are from the standard Italian dictionary of Nicola Zingarelli (Vocabolario della lingua italiana). “Falso … A agg.:… 2 Che è stato contraffatto, alterato con intenzione dolosa … SIN. Truccato. CONTR. Autentico.… 4 Che non è ciò di cui ha l’apparenza … SIN. Illusorio.… B s. m.… 3 Falsificazione, falsità … 4 Opera d’arte, francobollo, documento e sim. contraffatto.” “Falsificare … Contraffare, deformare, alterare con l’intenzione e la consapevolezza di commettere un reato.” “Falsificazione … 1 Atto, effetto del falsificare … SIN. Alterazione, contraffazione. 2 Documento o atto artificiosamente prodotto per sostutuire un originale perduto o guasto o per creare testimonianza dolosa.” “Contraffare … 2 Alterare la voce, l’aspetto e sim., spec. per trarre in inganno … 3 Falsificare.” “Facsimile … 1 Riproduzione esatta, nella forma della scrittura e in ogni particolare, di scritto, stampa, incisione, firma. 2 fig. Persona o cosa assai simile a un’altra.” “Pseudo- … primo elemento … che, in parole composte della terminologia dotta e scientifica, significa genericamente ‘falso’… In vari casi indica analogia esteriore, qualità apparente, semplice somiglianza puramente estrinseca, o qualche affinità con quanto designato dal secondo componente.” “Spurio … 1 Illegitimo … 2 Privo di genuinità, di autenticità.” “Apocrifo … 2 Detto di testo, spec. letterario, falsamente attribuito a un’epoica o a un autore. SIN. Spurio.”

2. See also Haywood (1987: 10–18).

3. For the terminology of this section see the chapter “Theory of Sign Production” in Eco (1979b).

4. Often a minimal material or formal variant serves to characterize the object as a unicum: two dollar bills of the same value are doubles as far as their use goes, but not from the bank’s point of view, since their serial numbers are different. Even in a case of perfect reproduction, the token that received the number first is considered theoretically “original.” Hence the interesting question whether we are to consider authentic a fake bill printed (with fraudulent intent) on authentic watermarked and security-threaded stock, with the plates of the Mint, by the director of the Mint in person, who assigns it the same number as another bill legally printed a few moments earlier. If it were ever possible to determine the priority of its printing, only the first bill would be authentic. Otherwise one would have to decide to arbitrarily destroy one of the two bills and consider the other the original.

5. The modern concept of the work of art as an unrepeatable unicum privileges its originality and its formal and material complexity, which, taken together, constitute the concept of authorial authenticity. Naturally in the practice of critics and collectors the notion of originality often prevails over the presence of relevant structural features. As a result, even a perfect copy of a statue, which reproduces, using the exact same materials, every aesthetically relevant feature of the original, is downgraded only because it is denied recognition of the privilege of originality. Problems of this sort crop up for the plastic and figurative arts but not for written texts, since any reproduction, be it printed or manuscript, of the same poetic text is assumed, for critical purposes, to be a perfect double of the original type (see the distinction between autographic versus allographic arts in Goodman 1968). They do, however, occur among bibliophiles, where in fact value is placed on the particular material consistency which renders one token (a copy of a rare book) something unique compared with other copies of the same book (evidence of possession, state of preservation, width of the margins, etc.).

6. A recent phenomenon is that of commercial facsimiles of precious illuminated manuscripts, in which the colors, the tactile feel of the gold leaf, the wormholes, and the transparency of the parchment are all reproduced with absolute fidelity, though the manuscript is not reproduced on real parchment but on paper (though it contrives to imitate the consistency of the original parchment). Even the reproduction of real parchment would display, when submitted to chemical tests, characteristics different from the antique original. And even if the reproduction were to be printed on recovered ancient parchment the same tests would demonstrate that the characters printed on it were made by mechanical means. And in any case the ancient parchment used for the reproduction would not be the original parchment of the manuscript.

7. In the same way we do not have false identification in texts written under a pseudonym when A (usually a famous person or someone otherwise known) produces O but would have it believed that it was produced by an unknown B (the identity of two objects is not an issue in such a case); in cases of plagiarism, in which B produces an object Ob which he presents as his own work, but using wholly or in part an object Oa produced by someone else (where B however does all he can to ensure that Ob will not be identified with Oa); or in cases of aberrant decoding, in which a text O, written according to a code C1, is interpreted as if it were written according to a code C2. Examples of this last practice are the oracular reading of Virgil as a Christian author in the Middle Ages, in the Baroque period the false interpretation of Egyptian hieroglyphics on the part of Athanasius Kircher, in modern times the reading of Dante as if he were writing in the secret code of the so-called sect of the Fedeli d’Amore (see Pozzato 1989). But in this kind of exercise there is no question of identification between two physical objects.

8. This continuing ascendency of logic over grammar in the thirteenth century was accurately described by Gilson (1952) in the chapter of his Philosophie au Moyen Age entitled “L’exil des belles-lettres.”

9. http://www.logicmuseum.com/authors/aquinas/superboethiumq2.htm.

10. In the De pulchro et bono, another case of false attribution—to Thomas Aquinas—and this time not just on the part of the Middle Ages but all the way down to our own century (see Chapter 8 in this volume). For a discussion of kalon, see the introduction by Pietro Caramello to the Marietti edition of the De divinis nominibus.

11. This of course is also the case with artistic fakes, like the fake Dutch masters painted fifty years ago by the extraordinarily talented contemporary artist Han van Meegeren.

12. See Grabmann (1906–1911, esp. Part IV of the first volume, devoted to the transmission of traditional knowledge), and Chenu (1950: 128–129). On how anthologies may give rise to a series of misunderstandings concerning originals that no one reads any more, see Ghellinck (1939: 95 and 105).

13. “Dicebat Bernardus Carnotensis nos esse quasi nanos gigantium humeris insidentes, ut possimus plura eis et in remotiora videre, non utique proprii visus acumine aut eminentia corporis sed quia quia in altum subvehimur et extollimur magnitudine gigantes” (“Bernard of Chartres used to say that we were like dwarfs seated on the shoulders of giants. If we see more and further than they, it is not due to our own clear eyes or tall bodies, but because we are raised on high and upborne by their gigantic bigness”), Metalogicon (1159) bk. 3, ch. 4. Translation from Henry Osborn Taylor The Mediaeval Mind ([1911]1919) vol. 2, p. 159. See Jeauneau (1967: 79–99) and Merton (1965).

14. An interesting link between Priscian and Bernard could be William of Conches, who mentions dwarves and giants in his glosses on Priscian’s Institutiones grammaticae. William’s text precedes that of John of Salisbury and was written in the years when William was chancellor at Chartres. But, while the first version of William’s glosses dates back to before 1123 (John’s Metalogicon is dated 1159), before Neckam, Peter of Blois, and Alain de Lille, all three cited by Merton, we find the aphorism in 1160 in a text from the school of Laon and later, around 1185, in the Danish historian Sven Aggesen. In the thirteenth century, the aphorism also appears in Gérard of Cambrai, Raoul of Longchamp, Gilles de Corbeil, Gérard of Auvergne, and, in the fourteenth century, in Alexandre Ricat, physician to the kings of Aragon, or other doctors like Guy de Chauliac and Ambroise Paré, as well as in Daniel Sennert. Gregory (1961) identifies it in Gassendi. Ortega y Gasset, in “Entorno a Galileo” (Obras completas V, Madrid 1947: 45), speaking of the succession of generations, says that men stand “one on the shoulders of another, and the one who is on top enjoys the impression of dominating the others, but he ought to realize that at the same time he is their prisoner.”

15. See, for example, the chapter on the spatial and temporal structures of the Middle Ages in Le Goff (1964).

16. See McGarry (1955: 167).