From the Tree to the Labyrinth: Historical Studies on the Sign and Interpretation

by Umberto Eco

  Where are we to look for the keys to decoding, since the text must after all be interpreted “correctly,” that is, according to an approved code? When he speaks about words, Augustine knows where to look for the rules—in classical grammar and rhetoric. But if Scripture speaks not only in verbis but in factis (De doctrina christiana II, 10, 15)—if there is, in other words, allegoria historiae in addition to allegoria sermonis (cf. De vera religione 50, 99)—then one must resort to one’s knowledge of the world.18

  Hence the resort to the encyclopedia, which traces an imago mundi, giving us the spiritual meaning of every worldly thing or event mentioned in Scripture. The Middle Ages inherited fascinating descriptions of the universe as a collection of marvelous facts from pagan culture: from Pliny to the Polyhistor of Solinus or the Alexander Romance. All they had to do was to moralize the encyclopedia, attributing a spiritual meaning to every object in the world. And so, following the model of the Physiologus, the Middle Ages began to compile its own encyclopedias, from the Etymologiae of Isidore of Seville to the De rerum naturis of Rabanus Maurus, to Honorius of Autun’s De imagine mundi or Alexander Neckham’s De naturis rerum, to the De proprietatibus rerum of Bartholomaeus Anglicus and the Specula of Vincent of Beauvais. The task was to provide, backed by the authority of tradition, the rules of correlation that would make it possible to assign a figural significance to any element in the physical world. And since authority has a nose of wax, and since every encyclopedist is a dwarf on the shoulders of the encyclopedists who went before him, they had no problem, not only in multiplying meanings, but in inventing new creatures and properties, that (on account of their curiouser and curiouser characteristics) would make the world into one immense speech act.

  At this point what is dubbed indifferently “medieval symbolism” or “allegory” takes separate paths. Separate at least in our eyes, which are looking for a handy typology, though these modes in fact interpenetrate continuously, especially when we consider that poets too will soon start writing allegorically like Scripture (see below what we have to say about Dante).

  We may distinguish, then, under the generic heading of symbolism (or the aliud dicitur aliud demonstratur), a series of different attitudes (Figure 3.1).

  Figure 3.1

  What we may call “metaphysical pansemiosis” does not interest us in the present context. This is the approach of Scotus Eriugena, for whom every element with which the world is furnished is a theophany that refers back to its first cause: “nihil enim visibilium rerum corporaliumque est, ut arbitror, quod non incorporale quid et intelligibile significet (De divisione naturae” (“there is nothing among visible and corporeal things that I can think of that does not signify something incorporeal and intelligent”) (De divisione naturae, 5, 3). Like the Victorines, Eriugena does not speak simply of the allegorical or metaphorical resemblance between terrestrial bodies and celestial things, but in particular of their more “philosophical” significance, which has to do with the uninterrupted series of causes and effects known as the Great Chain of Being (cf. Lovejoy 1936).

  Universal allegorism is that of the encyclopedias, bestiaries, and lapidaries: it represents a fabulous and hallucinatory way of looking at the universe, not for what it makes apparent but for what it might allude to: the difference, with regard to metaphysical pansemiosis, lies in the different philosophical awareness, in the metaphysical foundation, to be precise, of the ulterior meaning of sensible and corporeal things.19

  We have already spoken of scriptural allegorism and will do so again shortly; and what liturgical allegorism might consist of is intuitive.

  Poetic allegorism is that abundantly employed by secular poetry: Dante’s dark wood, say, or the whole of the Roman de la Rose. It imitates the modes of scriptural allegorism, but the facts presented are fictitious. If anything, oriented as it is toward moral edification, it may at most aspire to a cognitive function. But it is precisely in the case of the allegory of poets that a nexus of interesting problems comes to the fore.

  The Middle Ages abounds in allegorical readings of poetic texts (cf. De Bruyne 1946, I, 3, 8). Fables provide the first instance: naturally they speak of happenings that are patently false (talking animals and the like), though they do so with the intent of communicating a moral truth. If we read the various treatises that prescribe ways of correctly reading poetic texts (see, for instance, the Dialogus super auctores of Conrad of Hirsau), we will see that what they consist of are exercises in textual analysis. Faced with a poetic text, we must ask who is its author, what was the author’s purpose and intention, the nature of the poem or the genre to which it belongs, and the order and number of the books before going on to examine the relationship between littera, sensus, and sententia. As Hugh of Saint Victor observes in his Didascalicon, the littera is the ordered disposition of the words, the sensus is the obvious and simple meaning of the phrase as it appears at first reading, and the sententia is a more profound form of understanding, which can only be arrived at through commentary and interpretation.20

  All the authors insist on the primary need to examine the letter, expound the meaning of difficult words, justify the grammatical and syntactical forms, identify the figures and tropes. At this point one proceeds to interpret the meaning intended by the author, as this is suggested by the letter of the text. Then, to the hidden meaning, according to the formula aliud dicitur et aliud demonstratur. Now, it would appear that opinions differ concerning the distinction between sensus and sententia. For some interpreters analyzing a fable by Aesop or Avianus, the sententia would be the moral truth contained in the fable, according to which, in the fable of the wolf and the lamb, wolves are evil and lambs are good. But is this meaning, which the author makes so explicit beneath the integumentum or covering of the parable, the sensus or the sententia? For some interpreters, fables have a parabolic meaning, offered immediately to the reader, while the sententia would be a more deeply hidden allegorical truth, similar to that of the Scriptures (cf. De Bruyne 1946: 2:326–327).

  We have only to read Comparetti’s Vergil in the Middle Ages (1885) to see what sources the invitation to the Middle Ages to read the Roman poet allegorically came from. Medieval scholars may have been familiar with a commentary on Homer by Donatus that has since disappeared; they certainly knew Servius’s commentary and Macrobius’s observations on Virgil. Virgil was considered not only the greatest of poets (Homer was merely a legend, and his actual texts were unknown) but also the wisest of men. Accordingly, Bernard of Chartres, John of Salisbury, or Bernardus Sylvestris, among others, read the first six books of the Aeneid as a representation of the six ages of life. But what difference is there between this search for the epic’s allegorical sententia and the discovery of the parabolic meaning of a fable? The parabolic meaning seems to depend closely on the literal meaning, at a less subtle level than that of the allegorical sententia.

  Ulrich of Strasbourg (De summo bono I, 2, 9; cf. De Bruyne 1946: 2:314) says that fables, though they evidently say false things, can be taken as true, since the thing meant is not that conveyed by the words but by the sense that those words express. Alexander of Hales suggested adding to the four senses of Scripture (historical, allegorical, moral, and anagogical) the parabolic sense, which he reduced to the historical, distinguishing, however, within the historical sense, the sense secundum rem, in other words, the literal sense of the facts narrated, and that secundum similitudinem, as occurs in parables.21

  De Bruyne (1946: 2:312–313) attempts to systematize these differences between the various senses in the following way: the literal sense may be proper (or historical, in which an account is given of the actual events), figurative (typical, in the sense in which the individual represents the universal), parabolic and moral (in the secular sense, as in fables), or allegorical (or typical-figural, in factis); the spiritual sense, on the other hand, may be tropical (or moral) or anagogical.

  At this point, we should underscore a difference between the metaphorical sense (in which
the letter appears to be mendacious, unless we understand it to be figurative) and the moral sense of the fable, which could be ignored without the fable ceasing to signify things that are understandable, though considered false. But perhaps to the medieval mind it seemed equally false that a meadow could smile or that an animal could talk, only, in the first case, the falsity was in adjecto and, in the second, in the course of the events narrated. On the other hand, in these instructions for reading texts, the importance of identifying the metaphors is stated, but it does not seem that particular hermeneutical efforts are to be brought into play, whereas if one reads Aesop one must make an interpretive effort, however minimal, to understand the moral truth the author wished to express. The fact is we are faced with three different senses: (i) the sense of the metaphors, which, as we have seen, never poses a problem; (ii) the parabolic sense of the fables, in which we must indubitably attribute to the author a moralizing intent, if we are not to remain attached to the mendacious letter—and yet this moral sense is not obscure but evident; and (iii) the sense of the allegory, which allows us to know per speculum et in aenigmate.

  To make things still more complicated, we find in doctrinal circles impatience with the allegorical interpretation of secular poetry, and John of Salisbury, for instance, will say that, since humane letters must not draw a veil over sacred mysteries, it is ridiculous, harmful, and useless to look for anything beyond the literal sense (Polycraticus VII, 12).22

  This knot will be loosed, in exemplary but—to our way of thinking—astonishing fashion, by Thomas Aquinas.

  3.4. Metaphor in Thomas Aquinas

  Thomas (Summa Theologiae I, 1, 9) asks if the use of poetic metaphors in the Bible is permissible, and he seems to come to a negative conclusion, when he quotes the current opinion by which poetry is an infima doctrina or inferior teaching. And he seems to share this opinion when he says that “poetica non capiuntur a ratione humana propter defectus veritatis qui est in eis” (“human reason fails to grasp the import of poetical utterance on account of its deficiency in truth”) (Summa Theologiae II–II, 101. 2 ad 2). However, this affirmation should not be taken as a putdown of poetry or as a definition of the poetic in eighteenth-century terms as perceptio confusa. Instead, it is about recognizing poetry’s status as an art (and therefore of recta ratio factibilium or right judgment regarding things to be made), in which making is naturally inferior to the pure knowing of philosophy and theology.

  Thomas had learned from Aristotle’s Metaphysics that the efforts at storytelling of the earliest poet theologians represented a childlike form of rational knowledge of the world. In fact, like all Scholastics, he is uninterested in a doctrine of poetry (a subject for the authors of rhetorical treatises who taught in the Faculty of Arts and not in the Faculty of Theology). Thomas was a poet in his own right (and an excellent one at that), but in the passages in which he compares poetic knowledge with theological knowledge, he conforms to a canonical opposition and refers to the world of poetry merely as an unexamined alternative. He is impervious to the idea that poets can express universal truths, because he has not read Aristotle on the subject, and he therefore sticks to the received wisdom that poets recount fabulae fictae. On the other hand, he admits that the divine mysteries, which go beyond our ability to understand, must be revealed in allegorical form: “conveniens est sacrae scripturae divina et spiritualia sub similitudine corporalium tradere” (“Holy Scripture fittingly delivers divine and spiritual realities under bodily guises”) (Summa Theologiae I, q. 1, a. 9 co.). As for the reading of the sacred text, he specifies that it is based first and foremost on the literal and historical sense: when Scripture says that the Hebrews went out of the land of Egypt, it relates a fact; this fact is comprehensible and constitutes the immediate denotation of the narrative discourse. But the res, the things of which the sacred text supplies the record, were arranged by God as signs. The spiritual sense, then, is that meaning by means of which the things signified by the language refer to other things, and it is based accordingly on the literal sense. Thus, God disposes the same course of events, subject to his divine providence, to endow them with a spiritual meaning.23

  What we have here is not a rhetorical procedure, as would be the case with tropes or allegories in verbis; instead what we have are pure allegories in factis, in which it is the things themselves that act as signifiers of higher truths.24

  Up until this point Thomas would not have been saying anything new. But in his allusions to the literal sense he emphasizes a rather important notion, namely that the literal sense is quem auctor intendit. Thomas does not speak of a literal sense as the sense of the sentence (what the sentence says denotatively according to the linguistic code to which it refers), but rather as the sense attributed to the act of enunciation! Accordingly—we are interpreting Thomas’s words—if the sentence says that teeth are made of snow, we are not to understand that, grammatically speaking, the sentence expresses a mendacious proposition. The speaker’s intention, in using that metaphor, was to say that the teeth were white (like snow) and therefore the metaphorical construction is part of the literal sense, because it is part of the content that the speaker intended to say. In Super epistulam ad Galatas too, Thomas reminds us that both homo ridet and pratum ridet (figurative sense) are part of the literal sense (VII, 254). In III Sent. he says that in scriptural metaphors there is no falsehood (38, 1, ad 4).25

  In short, Thomas is prepared to speak of a secondary or spiritual meaning only when senses can be identified in a text that the author did not intend to communicate, and did not know were being communicated. And this is the case for an author (like the author of the Bible) who narrates facts without knowing that they have been prearranged by God as signs of something else.

  While we may speak, then, of a secondary sense of Scripture, things change when we move on to secular poetry or any other human discourse that does not concern sacred history. In fact at this point Thomas makes an important affirmation, which in a nutshell is this: allegory in factis is valid only for sacred history, not for profane history. God, so to speak, has limited his role as manipulator of events to sacred history alone, but we must not look for any mystic meaning after the Redemption—profane history is a history of facts not of signs: “unde in nulla scientia, humana industria inventa, proprie loquendo, potest inveniri nisi litteralis sensus” (“hence in no science discovered by human industry can we find, strictly speaking, anything beside a literal sense”) (Quaestiones quodlibetales VII q. 6 a. 3 co.).

  On the one hand this move—inspired by the new Aristotelian naturalism—calls into question universal allegorism, with its bestiaries, lapidaries, encyclopedias, the mystical symbolism of the Rhythmus alter, and the vision of a universe populated by entities at a high symbolic temperature. And naturally it sounds like an out-and-out repudiation of allegorical readings of the pagan poets. On the other hand, it tells us that when, in secular poetry, a rhetorical figure occurs (including metaphor) there is no spiritual sense, only a sensus parabolicus, which is part of the literal sense.26

  When, then, in the Scriptures Christ is designated through the figure of a goat (a scapegoat) what we have is not allegoria in factis, but a simple poetic procedure: allegoria in verbis. The poetic expression is not a symbol or an allegory of divine or future things, it simply signifies—parabolically and therefore literally—Christ (Quaestiones quodlibetales VII, 6, 15).27

  There is no spiritual meaning in poetic discourse or even in Scripture when they use rhetorical figures, because that is the meaning the author intended, and the reader easily identifies it as literal on the basis of rhetorical rules. But this does not mean that the literal level (as the parabolic and therefore rhetorical sense) cannot have more than one meaning. Which means in other words, though Thomas does not say as much apertis verbis (because the problem does not interest him), that there may be more than one level of meaning in secular poetry. Except that those different levels of meaning, couched in the parabolic mode, belong to the
literal sense of the sentence as understood by its enunciator. To the extent that, since the author of the Scriptures is God, and God can understand and intend many things at the same time, it is possible that in the Scriptures there are plures sensus or several meanings, even according to the merely literal sense.

  Likewise, we may speak of a simple literal meaning for liturgical allegory too, which employs not merely words but also gestures, colors, and images, since in that case the administrator of the rite intends to say something precise by means of a parable and we must not look, in the words that he formulates or prescribes, for a secret unintended meaning. Though the ceremonial precept, as it appeared in the old law, may have had a spiritual sense, when it was introduced into Christian liturgy it assumed a significance that was purely and simply parabolic.

  Thomas reorganizes a series of scattered notions and implicit convictions that explain why the Middle Ages paid so little attention to the analysis of metaphor. If what the author intended to say literally must be clearly understood through the trope, any attempt to create bold and unexpected metaphors would compromise their natural literalness. Medieval theory would not have been able to accept as a good metaphor or simile Montale’s bold comparison between life (and its travails and frustrations) and walking along a wall that has fragments of broken glass cemented on top of it, because the similarity had not been codified.28

  3.5. Dante

  Dante does not appear to pay the slightest attention to Thomas’s strictures (cf. Eco 1985). In Epistola XIII, explaining to Cangrande della Scala the keys for reading his poem, he says that the work is polysemos, that it has several senses, and he lists the four canonical levels—literal, allegorical, moral, and anagogical.29