by Umberto Eco
Aristotle declares that metaphors should be drawn from things that are not evident, just as in philosophy the sagacious mind recognizes, discovers, perceives (theorein) similarities between distant things (1412a 12). On the other hand, in 1405b he says that metaphors imply enigmas. When, apropos of the asteia (1410b 6 et seq.), he says that the poet calls old age kalámen or “a withered stalk,” he specifies that such a metaphor is productive of a knowledge (gnosis) through their common genus, inasmuch as both belong to the genus of things that have lost their bloom. Elegant enthymemes are those which help us learn in a new and rapid way and, in this as in other cases, the verbum cognoscendi used is manthanein, to learn. Those enthymemes are efficacious that are understood little by little as they are spoken and were previously unknown, or those we understand only at the end. In such cases we say that gnosis gínetai (“knowledge comes to be”). Moreover, the obvious metaphor, which is not at all striking, is rejected. When the metaphor makes us see things the opposite from the way we thought they were, it becomes evident that we have learned something, and our mind seems to say: “That’s the way it was, and I was mistaken about it.”
Metaphors, then, “put the thing before our eyes” (to poiein to pragma pro ommaton). This notion of “putting something before our eyes” is repeated several other times in the text, and Aristotle appears to insist on it with conviction: a metaphor is not a mere transfer but a transfer that is immediate in its evidence—but clearly unfamiliar, unexpected, thanks to which things are seen in action (1410b 34), or better, signified in action.
As for the many examples provided by the text, especially those that concern similes (1406b 20 et seq.), it is certainly difficult to say whether they may have sounded bold to the ears of Aristotle’s contemporaries, but all of them appear to be examples of original witticisms. The same can be said of the passage on the asteia (1411b 22). All the examples are provocative and so little used previously that they are attributed to a specific author. To call triremes painted millstones and taverns the mess-rooms of Attica is a fine way to show something in a new light.
But what is it that metaphor as a cognitive mechanism makes us see in a fresh light? Things themselves, or the way we were accustomed to seeing (and representing) things?
It appears that it is only in contemporary culture that we have realized that, in order to be understood, metaphors often require us to reorganize our categories. As Black (1979: 39–40) remarks, “some metaphors enable us to see aspects of reality that the metaphor production helps to constitute. But that is no longer surprising if one believes that the world is necessarily a world under a certain description—or a world seen from a certain perspective. Certain metaphors can create such a perspective.”35
Still, when Aristotle said that the invention of an effective metaphor “puts before our eyes” for the first time an unfamiliar relationship between two things, he meant that metaphor compels us to reorganize our knowledge and our opinions. Let us return to the Rhetoric (1405a) and the metaphor by which pirates are said to be purveyors or suppliers. Now, before the appearance of this metaphor there was nothing to associate an honest merchant who acquires, transports by ship, and resells his merchandise with a pirate who steals someone else’s merchandise. The astuteness of the metaphor consists in compelling us to identify a hierarchical organization of property that, on a lower level, distinguishes a violent action from a pacific one, but, on the higher level, lumps together genera and species of those who transport merchandise upon the sea. In this way the metaphor unexpectedly suggests a socially useful role for the pirate, at the same time leading us to suspect that there may be something not altogether above board about the transactions of the merchant. In this way, the categorical field becomes reorganized no longer on the basis of moral or legal considerations but on the basis of economic activity.
We have already remarked how, in seeking various explanations for the eclipse, Aristotle tried out various “ontologies” (and we are not going too far in using the term in the quintessentially modern sense we just recognized). Similarly, when, in On the Parts of Animals, he must decide, on the basis of empirical observations, which of the various biological phenomena are causes and which effects, Aristotle finds himself faced with the fact that ruminants (animals, that is, with four stomachs) have horns and lack upper incisors—with the embarrassing exception of the camel, which is a ruminant lacking upper incisors, but without horns.
Aristotle first proposes a definition whereby horned animals are animals which, since they have four stomachs—which makes internal rumination possible—have redirected the hard matter of the teeth into the formation of horns. In order to make the camel fit into this categorical organization, Aristotle must suppose that it did not need to redirect the hard matter into horns (because, being large, it had no need for further protection), but instead it deflected it to the gums and palate (Figure 1.19).
Figure 1.19
But why are ruminants the way they are? The fact that they are ruminants explains why they have horns, but having horns does not explain why they are ruminants. Faced with the need to define the category of ruminants, Aristotle puts forward the hypothesis that ruminants have deviated the hard matter from the mouth to the head for reasons of defense and have developed four stomachs as a consequence (Figure 1.20).
Figure 1.20
As we can readily see, these two definitions presuppose two different categorical organizations, in the first of which it is the fact of being a ruminant that determines the deviation of the incisors and makes possible the development of horns, in the other it is the deviation of the incisors for the purposes of defense that produced the four stomachs. The truth is that Aristotle, in suggesting a number of hypotheses regarding causes and effects, in no way attempts to construct pseudo-Porphyrian trees. He merely shows extreme flexibility in selecting as a genus what was previously a species and vice versa. In other words, he never tells us that the definition is based on an underlying ontological structure, rather what he does is to propose a methodology of division that makes an adequate definition possible. It is not the underlying tree that makes the definition possible, it is the definition that imposes an underlying tree, frequently ad hoc. But in his theory of metaphor Aristotle goes still further: he suggests that a creative and original use of language obliges us to invent a new ontology—and therefore, we might add, to enrich to some degree our encyclopedia.
Naturally, the new ontology is only valid as far as the comprehension of the creative text that imposes it is concerned. But we are entitled to suppose that, once the creative text has imposed a new ontology, however local, somehow or other it leaves a trace in our encyclopedia.
1.8.2. Joycean Ontologies
In my essay “The Semantics of Metaphor” (in Eco 1984c), a kind of reduced ontology was constructed, made up of all the expressions that appear in a certain section of Joyce’s Finnegans Wake, and an attempt was made to justify the various puns as passages among a series of phonetic, synecdochic, metonymic, or metaphoric associations.
The experiment was intended to demonstrate how, starting from whatever point of the textual universe one might chose as a sample, one could attain, by multiple and continuous pathways, as in a garden of forking paths, any other point.
In the schema presented in Figure 1.21,36 we may observe how the term Neanderthal evokes by phonetic association three other terms: meander, Tal (German for “valley”) and tale (“story,” in English), which combine to form the punning coinage cited in the book, meandertale. In the associative trajectory, however, intermediate nodes are created, provided by terms all of which appear in the text of Finnegans Wake and only there. At this point the associations may be phonetic or semantic in nature.
These interconnections demonstrate, moreover, how each term may become in its turn the archetype of an associative series that would lead us to identify, sooner or later, other associative chains. The whole diagram has a purely orientative value, in the sense that it reduces the ass
ociations numerically and dimensionally, but if we proceed from this ontology to Joyce’s text, we observe that all the associations registered by the network have been developed (in other words, we see how that ontology had its origin in the need to render explicit the associations that that text intended to provoke). Indeed, every association produces a pun that defines the book. The book is a slipping beauty (and hence a sleeping beauty who as she sleeps generates a series of lapsus through semantic slips, mindful of an error, etc.), a jungfraud’s messonge-book (where, to the associations already cited, that of message is added), a labyrinth (or meandertale) in which we find a word as cunningly hidden in its maze of confused drapery as a fieldmouse in a nest of coloured ribbons (the expression that gave rise to the schema, which naturally would have been clearer had the various circuits been colored differently). As a final synthesis, in the pages of the book the neologism meanderthalltale becomes the metaphorical stand-in for all that can be said about the book itself and which is said by the associative chains identified in the network.
Figure 1.21
In the associative sequences, semantic rather than phonetic in nature, the terms are associated through identity or similarity of properties (not real, but culturally imputed). If we reread the associative sequences we see that each of them could be constructed by referring to a “notional field” accepted in a given culture or to one of those typical linguistic carrefours or crossroads theorized by Trier, Matoré, and others. A survey of the notional fields acquired by a given culture would explain not only why Freud and Fraud may be connected, by phonetic similarity, but also Freud-dream and Freud-Jung.
Let us consider, for example, the sequence generated by Tal: on the one hand it refers us to space and place, genera of which Tal is, so to speak, the species. But space in turn refers to time since the relationship between space and time is a typical relationship of complementarity. The relationship between time and the past and between time and the cycles (corsi e ricorsi) of Giambattista Vico arises from a more or less textbook contiguity.
This means that on the one hand all of the connections were certainly culturalized before Joyce justified them by pretending to institute them or discover them, while on the other they become evident to us (and allow us to reconstruct the ontology underlying the text) only because Joyce brought them to light (by making obvious the relationships among the terms of a domain that he himself brought into focus).
What makes the pun creative is not the series of connections (which potentially precedes it because they are already culturalized): it is the decision to oblige us to construct, by way of an unfamiliar ontology, short circuits that are possible but not yet evident. Between message and dream there is no phonetic similarity and only a weak semantic contiguity (whereby, but only in certain cultures, or as part of a psychoanalytic koiné, a dream is a message), and to bring them together the reader has been obliged to make the leap over unconnected points of the diagram so as to get from songe to mensonge or from message to mensonge. But from that moment—from the moment when the text has spoken—those points are no longer unconnected.
Language, carrying to creative outcomes the encyclopedic process of unlimited semiosis, has created a new polydimensional network of possible connections. This creative “gentle violence,” once set in motion, does not leave unaffected the collective encyclopedia (and indirectly, not even the one shared by those who have not read Joyce). It has left behind a trace, a fruitful wound.
1.9. The Formats of the Encyclopedia
1.9.1. From the Individual to the Maximal
However intriguing this reconstruction of a “Joycean ontology” may have proved to be, it cannot be denied that the model of a reduced labyrinth illustrated in Figure 1.21 is infinitely more impoverished than Finnegans Wake taken as a whole. As useful as it has been in understanding a series of implicit and explicit connections at the basis of a number of puns, and as instructive as it perhaps is as a miniaturization of an encyclopedic network, nevertheless, just like the rest of the ontologies we have spoken of so far, it fatally reduces the riches of the Maximal Encyclopedia (of which the entire text of Finnegans Wake is in any case only a part) to which it certainly refers us, though by means of a work of domestication.
We remarked in section 1.4 that a Maximal Encyclopedia cannot be consulted in its entirety because it represents the sum total of everything that was ever thought or said, or at least of everything that could in theory be discovered, to the extent to which it has been expressed through a series of materially identifiable interpretants (graffiti, stelae, monuments, manuscripts, books, electronic recordings)—a sort of World Wide Web far richer that the one to which we have access through the Internet.
Pavel (1986) invited us to try a fascinating mental experiment. Let us suppose that an omniscient being is capable of writing or reading a Magnum Opus that contains all of the true assertions regarding both the real world and all possible worlds. Naturally, since we can speak of the universe using different languages, and since each language defines it in a different way, there exists a Maximal Collection (Pavel calls it the “Total Image”) of Magna Opera. Let us now suppose that God charges a number of angels with writing Daily Books for each individual human being, in which they take note of all the propositions (concerning the possible worlds of that individual’s desires and hopes and the real world of his acts) that correspond to a true statement in one of the books that make up the Maximal Collection of Magna Opera. The collection of Daily Books belonging to a given individual must be produced on the Day of Judgment, along with the collection of the Books that assess the lives of families, tribes, and nations.
But the benevolent genie who writes a Daily Book is not content to align true statements: he connects them, evaluates them, builds them into a system. And since on the Day of Judgment individuals and groups will each have a defending angel, the defenders will rewrite for each individual another astronomical series of Daily Books in which the same statements will be linked together in different ways, and differently compared to the affirmations of some of the Magna Opera.
Since infinite alternative worlds make up each of the infinite Magna Opera, the angels will write an infinite number of Daily Books in which affirmations that are true in one world and false in the other will be mingled together. If we further hypothesize that some of the genies may be clumsy and mix up affirmations registered as mutually contradictory by a single Magnum Opus, what we will end up with will be a series of compendiums, miscellanies, compendiums of fragments of miscellanies that amalgamate strata of books of different origins, and at that point it will be very difficult to say which books are truthful and which fictional, and with respect to what original. We will have an astronomical infinity of books each of which will straddle different worlds and we will no doubt consider as true stories that others have considered as fictional.37
This gives us a good idea of what the Maximal Encyclopedia might look like, if we substitute for the angels the human beings who took time to leave behind their traces (from the bison depicted in the Altamira caverns to the invention of writing and beyond). The legend Pavel narrates gives a reasonable representation of our situation when confronted by the universe of affirmations that we are accustomed to accept not as “true” but in any case enunciated.38
In the preceding sections we saw how, confronted with the virtual immensity of the Maximal Encyclopedia (a regulatory hypothesis, a stimulus to the understanding of sentences of every type), we usually attempt to reduce its format, to construct local representations with the purpose of understanding a single context. Nevertheless, this entire dialectic between local and global is not so simple. In other words, recognizing it does not mean answering a question but formulating one. When in a given context we endeavor to reconstruct the portion of encyclopedia probably activated by some enunciator, to what format of the encyclopedia are we referring? Clearly, if a child tells us that the sun has moved, in our understanding of what the child means we do not refer the s
tatement to complex cosmographical notions concerning a galactic revolution of the sun, but instead to the set of “ingenuous” habits of perception on the basis of which we say that the sun rises and sets. But what encyclopedic format do we refer to when we are talking to a scientist, to an educated person, to a farm laborer, to an inhabitant of a far-off country?
In Kant and the Platypus I discussed the difference between Nuclear Content (NC)—a set of interpretants on the basis of which both a lay person and a naturalist can agree on the properties evoked by the term mouse, both understanding in the same way the sentence there is a mouse in the kitchen—and Molar Content (MC), which represents the specialized knowledge that a naturalist may have of a mouse. We are justified, then, in thinking that on the one hand there is a Median Encyclopedia (shared in the present case by both the naturalist and the common native speaker) and on the other an unmanageable plethora of Specialized Encyclopedias, the complete collection of which would constitute the unattainable Maximal Encyclopedia. Accordingly, we could imagine the states (or strata) of what Putnam has called the social division of linguistic labor by hypothesizing a kind of solar system (the Maximal Encyclopedia) in which a great many Specialized Encyclopedias describe orbits of varying circumferences around a central nucleus (the Median Encyclopedia), but at the center of that nucleus we must also imagine a swarm of Individual Encyclopedias representing in sundry and unforeseeable ways the encyclopedic notions of each individual.