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

by Umberto Eco

  9.6. Thomas Aquinas

  In his commentary on the De interpretatione, Thomas Aquinas, who remains faithful to Aristotle’s positions, after distinguishing the first operation of the intellect (perception) from the second (“scilicet de enunciatione affirmativa et negativa”), defines interpretatio as “vox significativa quae per se aliud significat, sive complexa sive incomplexa” (“significant vocal sound—whether complex or incomplex—which signifies something by itself”) (Proemium 2). But immediately afterward he makes it clear that nouns and verbs are merely “principles” of interpretation, which is to be identified exclusively with the oratio, that is, with all those propositions “in qua verum et falsum inveniuntur.”

  At this point he uses significare for the nouns and verbs (I, ii, 14), as well as for those voces that signify naturally, such as the moaning of the sick and the noises made by animals; but, as far as human voices are concerned, they do not immediately signify the things themselves but the general concepts, and only “eis mediantibus” (through them) do they refer to singularia (I, 2, 15).

  He later states that the name signifies its definition (I, ii, 20). True, when Thomas speaks of composition and division, that is, of affirmation and negation, he says the former “significat … coniunctionem,” while the latter “significat … rerum separationem” (I, iii, 26), but it is clear that even here (where language refers to what is or is not the case) what is signified is an operation of the intellect. It is only the intellect, whose operations are signified, that may be defined as true or false with respect to the actual state of things: “intellectus dicitur verum secundum quod conformatur rei” (I, iii, 28). An expression can be neither true nor false, it is merely the sign that significat a true or false operation of the intellect.5

  The verb denotare, in all of its various forms, occurs 105 times in the Thomistic lexicon (to which we may add two occurrences of the noun denotatio), but it appears that Thomas never used it in the strong extensional sense, in other words, he never used it to say that a given proposition denotes a state of affairs, or that a given term denotes a given thing.6

  It is occasionally used with the sense of “to signify metaphorically or symbolically that …” See, for instance, the commentary In Job 10, where it is stated that the roaring of the lion stands for Job (“in denotatione Job rugitus leonis”). There is an ambiguous passage in III Sent. 7, 3, 2, which says: “Similiter est falsa: ‘Filius Dei est praedestinatus,’ cum non ponatur aliquid respectu cujus possit antecessio denotari.” But it could be argued that what Thomas is talking about in this case is the mental operation that leads to the understanding of a temporal sequence.

  9.7. Suppositio

  Authors like Boethius, Abelard, or Thomas Aquinas, more concerned with the problem of signification than with that of denomination, were primarily interested in the psychological (today we would say “cognitive”) aspects of language. Certain of our contemporary scholars, however, committed to the rediscovery of the first medieval manifestations of a modern truth-conditional semantics, find the whole question of signification to be a very embarrassing problem, upsetting as it does the purity of the extensional approach, firmly established apparently by the theory of suppositio.7

  In its most mature formulation, supposition is the role a term, once inserted into a proposition, assumes so as to refer to the extralinguistic context. The road, however, that leads from the first vague notions of suppositum to the more elaborate theories like that of Ockham is long and winding. De Rijk (1962–1967, 1982, n. 16) has traced the path by which, in discussing the relationship between a term and the thing to which it refers, the notion of signification (understood as the relationship between words and concepts, or species, or universals, or definitions) becomes ever less important.

  We may observe how, for instance (De Rijk 1982: 161 et seq.), the disciples of Priscian spoke of names as signifying a substance at the same time as a quality (a formula in which the latter no doubt represented the universal nature of the thing and the former the individual thing), so that as early as the twelfth century we find the verb supponere as the equivalent of significare substantiam, in other words, signifying the individual thing. It is true, however, that authors like William of Conches insist that names do not signify either substance or quality or a thing’s actual existence, but only its universal nature, and that during the twelfth century the distinction is maintained between signification (of concepts and species) and denomination (the denotation of concrete individual things—see, for example, the Ars Meliduna).

  It is, however, clear how, little by little, in the fields of logic and grammar, the cognitive is superseded by the extensional approach, and how “in successive phases, the real meaning of a term became the focus of general interest, with the consequence that reference and denotation became far more important than the over-abstract notion of signification. What a term signifies first and foremost is the concrete object to which it can correctly be applied” (De Rijk 1982: 167).

  Notwithstanding this development, this novel point of view is not usually expressed using terms such as denotatio, whose semantic domain remains ill-defined.8 Peter of Spain, for example, uses denotari in at least one passage (Tractatus VII, 68), in which he states that, in the expression sedentem possibile est ambulare (“to someone seated ambulating is possible”), what is denoted is not the concomitance between being seated and ambulating, but that between being seated and having the possibility (potentia) to ambulate. Once again, it is difficult to say whether denotare has an intensional or extensional function. Furthermore, Peter considers significare in an extremely broad sense, given that “significatio termini, prout hic sumitur, est rei per vocem secundum placitum representatio” (Tractatus VI, 2), and it is impossible to decide whether this res is to be considered as an individual or a universal (De Rijk 1982: 169).

  On the other hand, Peter does introduce an honest-to-goodness extensional theory simply by developing a notion of suppositio distinct from that of signification (see also Ponzio 1983, who has an interesting reference to Peirce, CP 5.320): what Peter says in fact is that suppositio and significatio are different in that the latter is concerned with the imposition of a vox to signify something, while the former is the meaning of the same term (which already in and of itself and in the first instance signifies that given thing) inasmuch as it stands for something particular.9

  In Peter’s theory, however, there is a difference between standing extensionally for a class and standing extensionally for an individual. What we have in the first case is a natural supposition (suppositio naturalis), and in the second an accidental supposition (ibid., 4). Along the same lines, Peter distinguishes between suppositio and appellatio: “differt autem appellatio a suppositione et a significatione, quia appellatio est tantum de re existente, sed significatio et suppositio tam de re existente quam non existente” (Tractatus X, 1).

  De Rijk (1982: 169) affirms that “Peter’s natural supposition is the exact denotative counterpart of signification.” To be sure, we may insist that homo signifies a certain universal nature and supposes all (possible) existing human beings or the class of humans. What Peter does not say, however, is that homo signifies all existing human beings or that it denotes them, though the entire question does not substantially change.

  Up to this point, the terminological landscape that lies before us is still somewhat confused, considering that each of the technical terms considered so far covers at least two different domains (except for “denotation” and “designation,” which are still more indeterminate). This is illustrated by the diagram in Figure 9.5.

  A significant change occurs with William of Sherwood, who “unlike Peter and the majority of 13th-century logicians … identifies the significative character of a term with its referring exclusively to actually existing things” (De Rijk 1982: 170–171).

  Figure 9.5

  This will be the position of Roger Bacon, for whom signification becomes denotative in the modern extensional use of th
e term—despite the fact that he never employs a term such as denotatio.

  9.8. Bacon

  In his De signis (Fredborg et al. 1978, hereinafter DS), Bacon sets up a relatively complex classification of signs (fundamentally confirmed in other works by the same author, such as the Compendium studii theologiae), which presents a number of elements of interest to the semiotician. This classification has already been discussed,10 and we saw that Bacon employs the terms significare, significatio, and significatum in a sense radically different from the traditional one.

  In DS II, 2, he states that “signum autem est illud quod oblatum sensui vel intellectui aliquid designat ipsi intellectui.” A definition of this kind might appear similar to that of Augustine—but only if we understand Bacon’s “designat” as the equivalent of Augustine’s “faciens in cogitationem venire.” We must, however, point out two considerations that differentiate Bacon from Augustine. First of all, “oblatum sensui vel intellectui” implies that Bacon assumes a less radical stance than Augustine via-à-vis the sensible qualities of signs, given that he repeatedly admits that there may also be intellectual signs, in the sense that concepts too may be considered to be signs of things perceived. In the second place, for Augustine the sign produces something in the mind, while for Bacon a sign shows something (that exists outside of the mind) to the mind.

  Therefore, for Bacon signs do not refer to their referent through the mediation of a mental species, but are directly indicated, or posited, to refer immediately to an object. It makes no difference whether this object is an individual (something concrete), a species, a sentiment or a passion of the soul. What matters is that between a sign and the object that it is supposed to name there is no preliminary mental mediation. The mind steps in, so to speak, after the fact, to register the designation that has already taken place. As a result, Bacon uses significare in an exclusively extensional sense.

  It should be borne in mind, however, that Bacon distinguishes natural signs (physical symptoms and icons) from signs “ordinata ab anima et ex intentione animae,” in other words, signs produced by a human being with some purpose in mind.

  Among the signs ordinata ab anima are words and other visible signs of a conventional nature, such as the circulus vini or barrel hoop that taverns displayed to identify themselves, and even the goods displayed outside shops, inasmuch as they signify that other members of the same class to which they belong are on sale within. In all of these cases Bacon speaks of impositio, that is, of a conventional act by means of which a given entity finds itself having to stand for something else. Clearly, for Bacon convention is not the same as arbitrariness: the merchandise on display is chosen conventionally but not arbitrarily (the objects act as a kind of metonymy, the member of the class for the class as a whole). The circulus vini too is designated as a sign in a conventional and nonarbitrary manner, inasmuch as it points to the hoops that hold the barrels together, and acts simultaneously as both synecdoche and metonymy, representing a part of the container that holds the wine ready to be sold.

  But in DS most of the examples are taken from verbal language and hence, if we wish to follow Bacon’s line of thought, it would be better not to stray too far from what is probably the paramount example of a system of conventional and arbitrary signs.

  Bacon, however, is not so naïve as to say that words signify exclusively individual and concrete things. He contends that they name objects, but these objects may also exist in a mental space. Signs in fact can also name nonentities, such as infinity, a vacuum, the chimera, and nonbeing itself (DS II, 2, 19; see also II, 3, 27, and V, 162).

  This implies that, even when words signify species, this occurs because they point extensionally to a class of mental objects. The relationship is always extensional, and the correctness of the reference is guaranteed only by the actual presence of the object signified. A word is truly significant if, and only if, the object it signifies is the case—if nothing else if it is the case that it is thought.

  Admittedly, Bacon says (DS I, 1) “non enim sequitur: ‘signum in actu est, ergo res significata est,’ quia non entia possunt significari per voces sicut et entia,” but this position cannot be equated to Abelard’s insistence that even an expression like nulla rosa est signifies something. In the case of Abelard rosa was significant insofar as significare was considered from an intensional point of view, and, within this framework, the name signified the concept of the thing, even if the thing did not exist or had ceased to exist. Bacon’s position is different: when one says “there is a rose” (and when there being a rose is the case), the meaning of the word is given by the actual concrete rose, but when one makes the same affirmation and no such rose exists, then the word rose does not refer to an actual rose, but to the image of the supposed rose that the speaker has in mind. There are two different referents, and in fact the sound rose itself is a token of two different lexical types.

  Let us weigh carefully the following passage. Bacon states that “vox significativa ad placitum potest imponi … omnibus rebus extra animam et in anima,” and he admits that we may name conventionally both mental entities and nonentities, but he insists on the fact that it is impossible to signify with the same vox both the individual object and the species. If, to name a species (or any other mental passion), one intends to use the same word already used to name the corresponding object, we must set in motion a secunda impositio (DS V, 162).

  What Bacon intends to clarify is that, when we say “homo currit” (“the man is running”) we do not use the word homo in the same sense as in the sentence “homo est animal” (“man is an animal”). In the first case the referent of the word is an individual, in the second a species. There are then two equivocal ways of using the same expression. When a potential customer sees the barrel hoop advertising wine in a wine shop, if there is wine, then the hoop signifies the actual wine. If there is no wine, and the customer is misled by a sign that refers to something that is not the case, then the referent of the sign is the idea or image of wine that has taken shape (erroneously) in the customer’s mind.

  For the people who know there is no wine, the hoop has lost its ability to signify, in the same way in which, when we use the same words to refer to things in the past or the future, we do not use them in the same sense as we do when we refer to actual things that are present. When we speak of Socrates, referring, that is, to someone who is dead, and express our opinions about him, in reality we are using the expression Socrates with a new meaning. The word “recipit aliam significationem per transsumptionem,” it is used in an ambiguous way compared with the meaning it had when Socrates was alive. “Corrupta re cui facta est impositio, non remanebit vox significativa (DS IV, 2, 147). The linguistic term remains, but (as Bacon remarks at the beginning of DS I, 1) it remains only as a substance deprived of its ratio and of the semantic correlation that made its material occurrence a word.

  In the same way, when a child dies, what is left of the father is the substantia, not the relatio paternitatis (DS I, 1, 38).

  When we speak of individual things, “certum est inquirenti quod facta impositione soli rei extra animam, impossibile est (quod) vox significet speciem rei tamquam signum datum ab anima et significativum ad placitum, quia vox significativa ad placitum non significat nisi per impositionem et institutione,” while the relationship between the mental species and the thing (as the Aristotelian tradition was also aware) is psychological and not directly semiotic. Bacon does not deny that species can be the signs of things, but they are so in an iconic sense: they are natural signs, and not signs ordinata ab anima. The vox thus signifies only the individual thing and not the species (DS V, 163). As has already been demonstrated, when we decide to use the same term to name the species, what we have is a second imposition.

  Bacon subverts, then, once and for all the semiotic triangle implicitly formulated since Plato, according to which the relationship between words and referents is mediated by the idea, the concept, or the definition. At this juncture,
the left-hand side of the triangle (the relationship, that is, between words and concepts) is reduced to a merely symptomatic phenomenon (see Figure 9.6).

  Figure 9.6

  In Chapter 4, on the barking of the dog, we raised the question of whether Bacon had relied on Boethius’s translation of De interpretatione 16a, in which both symbolon and semeion were translated into Latin with the same word, nota, or whether he might not have gone back to the original, concluding from it that words are first and foremost in an exclusively symptomatic relationship with the passions of the soul. Accordingly, he interprets (DS V, 166) the passage in Aristotle from his own point of view: words are essentially in a symptomatic relation with species, and at most they can signify them only vicariously (secunda impositio), while the only real relation of signification is that between words and referents. He disregards the fact that, for Aristotle, words were, so to speak, symptoms of the species with reference to a temporal sequence, but that in any case they signified the species, to the point that we can only understand things named through the mediation of species already known.

  For Aristotle, and in general for the medieval tradition prior to Bacon, extension was a function of intension, and in order to ascertain whether something was in fact the case, one had first to understand the meaning of the statement. For Bacon, on the other hand, the meaning of the statement is the fact of which the referent is the case.

  What is of most interest to Bacon is the extensional aspect of the entire question, and this is why the relationship of words to what is the case looms so large in his treatise, while the relationship of words and their meaning becomes at best a subspecies of the referential relationship.

  We can thus understand why, in the context of his terminology, significatio undergoes a radical transformation from the meaning it had had until now. Before Bacon, nominantur singularia sed universalia significantur, but with Bacon and after him significantur singularia, or at least significantur res (though a res may be a class, a sentiment, an idea, or a species).