Umberto Eco and Charles Peirce: A slow and respectful convergence

1 Introduction

Umberto Eco took many of his semiotic concepts from Peirce, beginning in the 1960s and increasing his Peircean inspiration over the course of the years. At first glance, the two seem to have little in common.

While Peirce dedicated his life “To erect a philosophical edifice that shall outlast the vicissitudes of time” (CP 1.1), where semiotics is placed in an organized classification of science (see Proni 1992), Eco never wanted to give his theory of the sign an explicit philosophical foundation. This does not mean Eco has no implicit philosophy. And he appears to be more inclined towards nominalism than realism. At the same time, Eco made his intellectual debut taking a stand against idealistic aesthetics, whose main Italian representative was Benedetto Croce, the dominant figure in Italian philosophy in the first decades of twentieth century (Eco 1962; Proni 1988, Proni 1998). Since Peirce professed on several occasions to be a realist (a scholastic or Scotist realist, to be precise) ^[1] and an idealist (close to Schelling’s position; CP 4.453, 5.102), it seems that the two belong to opposite philosophical fields.

It would require too many pages to explain why, over three decades, Eco has taken so much from Peirce but not his philosophical approach. This choice, however, left Eco with a weak foundation for his semiotics, and he became progressively aware of this lack, particularly with regard to the problem of semiosis in perception.

In the 1960s Eco can still be defined as a structuralist, though he is very clear in denying an immanent character to any structure (see Eco 1968a). Then he progressively comes closer and closer to Peirce’s positions, trying at the same time to maintain two points: the conventional nature of signs and the central role of codes. The first implies the second, and the role of codes is later reduced to a particular case in the more general theory of the interpretant (Eco 1984a).

Eventually, in Kant e l’ornitorinco (1997) [Kant and the Platypus, 1999], Eco showed he had found in Peirce an approach to the foundation of semiotics he might embrace without surrendering to metaphysical bents, namely, phenomenology. In fact, Eco had been intellectually raised in an existentialist and phenomenological environment, and it was natural for him to adopt such a stance. Nonetheless, even now he still does not take an explicit philosophical position.

In Kant and the Platypus, then, the nature of the sign remains conventional, but a kind of “basic semiosis” is introduced in order to have a starting point for interpretation or decoding. Basic semiosis is, Eco writes “a message without a code” (1999a: 383, note 35). Thus implying that there are non-conventional signs.

Therefore Eco, eventually, though in an indirect way, adopts Peirce’s phenomenological view, but he still refuses to go beyond the empirical, ordinary sensorial representations that he calls “alfa mode” (Eco 1999a: 383). To be without a code a message must have no division between content and expression, since the code is the relation between them. In Peirce this happens only with Qualities of Feeling, but they have a particular mode of being or ontological status: that of Possibility or Firstness. And Firstness appears (we cannot say “exists” since Qualities of Feeling “seem” but don’t “exist”) before the distinction between the inner and the outer world, the Self and Not Self, arises in consciousness (CP 1.354–1.368, 5.41–5.56). Eco respects the division between semiotics and philosophy: he does not want to take ontological positions. In 1999, however, his approach to Peirce’s theory is almost complete. With the acceptance of basic semiosis, the notion of sign is founded on a non-conventional basis, and the central role of code is abandoned in favor of an interpretive explanation.

Thus, the relation between Eco and Peirce can be outlined as a progressive convergence. Let us now try to see in more detail the history of this slow but steady discovery and acknowledgment.

2 1968: the inquiry on codes

Umberto Eco’s work on semiotics began within the debate on structuralism, which was vibrant in the European cultural landscape of the 1960s. His contribution appeared in the book, La stuttura assente (1968) ^[2] [The Absent Structure] with the subtitle Introduzione alla ricerca semiologica [Introduction to semiotic research]. ^[3] Eco had encountered French structuralism, Roman Jakobson and Russian formalism some years before. The book that launched him onto the Italian intellectual scene was Opera aperta (1962) [The Open Work], which focused mainly on the aesthetics of interpretation with its starting point in the work of Luigi Pareyson, the supervisor of his dissertation. Already at this time, the central role of interpretation in aesthetics and semiotics was clear to Eco and when he discovered Peirce (through Jakobson, Morris, and Ogden and Richards), he found a confirmation and, moreover, a semiotic treatment of the interpretive process that was a more logically and philosophically correct description.

In La struttura assente (1968a) [The Absent Structure], Eco comes up with an original synthesis between the semiology of French origin and Peirce’s and Morris’ semiotics. The structuralist influence is still strong, though moderated by a pragmatic stance (Eco 1968a: 360 and Section D), and Peirce’s theory of the interpretant is explicitly presented (1968a: 35–36). The theory is described only in brief, yet Eco clearly sees in the Peircean notion of interpretant a productive possibility for his semiotics. Since, as Eco writes, the interpretant is, in Peirce’s view, “another representation that refers to the same object” ^[4] (1968a: 36, my translation), then the only way to establish the interpretant of a sign is by means of another sign, and so on, in a repetitive series called “unlimited semiosis,” a key concept developed by the American philosopher. Interpretation will thus be capable of self-reflection, “language would then be a system that clarifies itself, by means of successive systems of conventions that explain each other” (1968a: 36). In 1968, Eco is still reluctant to admit that the theory of interpretation can remove the need for a metalanguage in order to talk about meaning; he just notices that this way the code explains itself and that “since a language that talks about language is a metalanguage, then semiology brings about a hierarchy of metalanguages and nothing else” (1968a: 37).

At this stage Eco’s semiotics pivots on the notion of code. In the analysis of cinema, advertising, architecture, the keys are always the visual codes that allow one to grab the more or less hidden meanings. However, in codes, Eco is less interested in the structure of codes than in the continuous change produced by artistic and communicative practices (1968a: 97).

3 1975: the supremacy of interpretation

La struttura assente [The Absent Structure] marks the beginning of a semiotic inquiry that continues to develop with Le forme del contenuto (1971) [The Form of the Content] e Il segno (1973) [The Sign], assuming a definitive form in Trattato di semiotica generale (1975a) [A Theory of Semiotics, 1976a], in which Peirce’s theory of interpretation plays a fundamental role. A whole chapter is dedicated to its explanation (1976a: Ch. 2.7).

Eco adopts, with some changes, Peirce’s definition of sign, ^[5] “I propose to define as a sign everything that, on the grounds of a previously established convention, can be taken as something standing for something else” (1976a: 16). The change in Peirce’s theory seems to be a minor detail, yet it carries much weight. While, in Peirce, convention is only one of the three relations a sign can have to its object (“what it stands for”), Eco defines the sign function, following Saussure, only as a conventional relation. This allows him to preserve the central role of code, “The process of signification takes place only when a code exists.” ^[6] The notion of code, however, becomes in Eco something very different from Saussure and Hjelmslev.

First, Eco looks for a more specific definition of code. In this search, he criticizes, one by one, the different semantic models for their naturalistic or ontological biases (1976a: 126), and ends with the Global Semantic Universe or Model “Q” (1976a: 121), a “working hypothesis” not entirely representable and based on “a process of unlimited semiosis” (1976a: 122). In fact, since the code is not an a priori entity but an explicative hypothesis, then it is the product of semiotic activity and not its condition, and because such activity is based on interpretation, the code is just the trace of all the interpretive acts of a given culture, namely, the perpetual flow of semiosis.

Interpretive activity, though, does not end once a code is produced. A successful model must also account for its use and continuous change. Interpretation has the logical form of inference. The theory of inference or argument is a keystone not only of Peirce’s semiotics (or logic, Peirce, following Locke, identified logic with semiotics; see CP 2.227) but for his whole philosophical system. In Peirce, the Universe itself is a sort of gigantic argument unrolling itself in infinite evolution towards final logical perfection. Semiotically speaking, inference, according to Peirce, is a kind of sign (CP 2.252). Inference plays a role in establishing conventions (Eco 1975a: 29) and in interpreting uncodified expressions. Thus Eco presents the Peircean notion of abduction, the form of reasoning introduced in logic by Peirce (Eco 1975a: 185). ^[7] The inferential nature of semiosis will become from now on more and more clear in Eco’s theory: interpretation is always an inferential act, what changes is only the degree of automatism and the certainty of conclusion.

Interpretation thus becomes a more general concept than code, because it does not explain only the decoding of expressions but also the production of sign-functions (Eco 1975a: 185–187). Eco is careful, as Peirce himself was, to leave no room for mentalism, and he quotes one of Peirce’s definitions of the interpretant as “that which the sign produces in the quasi-mind which is the interpreter; but it can also be conceived as the definition of the representamen (and therefore its intension)” (Eco 1976a: 68). Eco then defines the interpretant as “another representation which is referred to the same ‘object’.” (1976a: 68). Such representations can be of five different kinds: equivalent sign-vehicles; indexes directed to an occurrence or quantifiers; definitions; emotive associations with connotative value; translations into another language (Eco 1976a: 70). Now unlimited semiosis is seen in a more assertive way as an alternative to metalanguage: “Semiosis explains itself by itself … this continual circularity is the normal condition of signification and even allows communication to use signs in order to mention things” (Eco 1976a: 71).

Yet, although Eco admits that the notion of interpretant is more comprehensive, he does not abandon the idea of mixing it with that of code. For this reason he builds a restricted definition of interpretant, capable of functioning within the framework of the code theory, limited to three semantic categories: a) synonymy; b) componential representation of a sememe; c) seme or semantic component (Eco 1976a: 72).

In A Theory of Semiotics interpretation also plays a fundamental role in the theory of mentioning or referring. Eco does not oppose an inferential conception of signification to an intuitive notion of perception but sees the “problem of perception as interpretation of sensory disconnected data which are organized through a complex transactional process by a cognitive hypothesis based on previous experience” (1976a: 165). Perception is thus a semiosic process, a representation, a sign, and concepts are also signs (CP 5.264–317). Eco refers not only to Peirce but also to the psychology of perception of Jean Piaget, to the nominalistic British tradition, and to Edmund Husserl. “The object,” Eco writes, “is not necessarily a thing or a state of the world: it is rather a rule, a law, a prescription” (1979a: 29, my translation). ^[8] However, as with any constructive theory of perception, the problem is only moved, not solved. There must always be some first elements of knowledge that are not produced by the subject, otherwise there is no knowledge of the external world. And how do these elements account for the external world? These questions will only find an answer in Eco several years from now.

4 1979: towards a pragmatics of texts

In A Theory of Semiotics Eco adopted Peirce’s theory of interpretation, in Lector in fabula (1979a) ^[9] he extended the theory to textual semiotics. Eco starts from Peirce’s division between Dynamical and Immediate Object or Ground. ^[10] The Dynamical Object is the “really efficient but not immediately present Object” (CP 8.343), that to which the Sign directs, while the Immediate Object is “the Object as the Sign represents it” (CP 8.343). Interpretant and Immediate Object or Ground are, in Peirce, synonyms of meaning (although with some differences). One of the peculiar features of Peirce’s semiotics is that meaning is not a central concept. The term itself is used in its common sense and does not appear in the definitions of representation or Sign. Of course, Peirce has a theory of meaning, in the sense in which linguistics and semiotics have considered it in the twentieth century, but it has to be framed in the triadic model, and not in the expression-content (or biplanar) model. In extreme synthesis, we might say that each of the three elements of Representation possesses some properties of meaning. Immediate Object is the significant part of the Sign (being the Representamen the mere sign-vehicle), that is, the portion of the thing or event that carries the information (see Eco 1979b: 183).

For example, in a weathercock, the Immediate Object is the direction it assumes, while its color, smell, and other features have no semiotic function. This is only a part of its meaning. The Interpretant takes into account the Immediate Object and assumes that it stands for its Dynamical Object: “This weathercock shows that the wind blows form the North.” Thus, the interpretant means here and now in its pragmatic signification, that is, the full meaning, which includes context and its circumstances. The Dynamical Object is that which puts the whole process in motion, standing behind the scenes, unreachable in its completeness (as the Ding an Sich in Kant) but effective in its empirical existence. In interpretation it is that part of meaning that has to do with the infinite approach to the referent, its “unveiling.” To stick to our example, the rotation of the weathercock can convey another trait of its Object (the wind), namely, that it is inconstant.

If we look at the other side of interpretation, namely, to sign production, the Object must be in the background when we progressively add to the sign more traits of the Referent. For example, I may add to the weathercock a rotating wheel to measure, besides direction, the speed of wind.

Another idea of Peirce that was taken up by Eco is that a term (or rheme) is a potential proposition, and a proposition (or dicisign) a potential argument. This allows him to say that “the meaning of a sign inchoatively contains all the texts within which that sign can be inserted” (1979b: 184). From this point he develops one of the central theses of the book, “a sememe is in itself an inchoative text, whereas a text is an expanded sememe” (1979b: 175). This conclusion represents an extension with respect to Peirce. While Peirce limited the expandability of term and proposition to their logical consequences, Eco added the possibility of expansion on the basis of discoursive implications, produced by interpretations sedimented in the Encyclopedia (or Global Semantic System or Model Q). In doing this, Eco corrects his own definition of the interpretant in a less “lexicalist” direction, “we must not forget that, for Peirce, not only a word or an image are signs, but also a proposition or a whole book” ^[11] (1979a: 35, my translation).

In Lector in fabula Eco raises, for the first time, the problem of the limits of interpretation, which will assume more and more relevance in his later research. The question, synthetically put, is that since the interpretative process takes the form of unlimited semiosis, “This infinite series could, however, make the semantic encyclopedia unattainable and the work of semantic analysis continuously baffled by its own need of completedness” (1979b: 189). It is thus necessary to somehow contain the interpretive fugue. According to Peirce, the flow of interpretation originated by a sign takes a rest (it does not die, because it can always be revived) with the formation of conscious habits, namely, norms for action, called Final Interpretants (see Peirce CP 5.11–13; 5.464–494). This is the core concept of pragmatism, defined by Peirce as “a method of ascertaining the meanings of hard words and of abstract concepts” (CP 5.464) by developing their conceivable consequences.

Eco initially acknowledges two limits to interpretation. The first is the universe of discourse itself. In ordinary discourse, in fact, we do not consider an infinite number of semantic implications (though it would be theoretically possible) but we limit the encyclopedic representation in such a way as to be both logically acceptable and viable (Eco 1979a, Eco 1979b: 88). The second is Peirce’s pragmatic limit. Eco poses the problem, which never troubled Peirce, about which habit a literary text might bring about, considering that it is bound to be different to that which the interpretation of a chemical definition, consisting of a series of operative instructions, requires. ^[12] One hypothesis is that some literary texts produce “messages,” in the sense of moral syntheses. Eco’s example is Le rouge et le noir by Stendhal, which “prescribes what one must do in order to acquire the practice of acting and modifying the world” ^[13] (1979a: 49, my translation). There are obviously more possibilities. The final interpretant of an aesthetic text can be the request for further reading moved by the mere pleasure of fruition, a change in introspective attitude, the choice of correcting one’s behavior in a sometimes radical manner, and so on. Certainly here individual competences and local encyclopedias also come into play.

Extending pragmatism to cover the case of literary texts raises another question, namely, the testing of interpretations. In fact, according to Peirce, habits produced by interpretation are selected by interaction with the environment, following the theory of evolution. Yet, this happens only with habits that imply such a selective interaction “outside of these it is probably of more advantage to the animal to have his mind filled with pleasing and encouraging visions, independently of their truth; and thus, upon unpractical subjects, natural selection might occasion a fallacious tendency of thought” (CP 5.366). The pragmatic approach to text does not necessarily produce an intellectual aesthetics. Acquiring pleasant sensations can be as useful as logical correctness of thought, and that is without considering that a hedonistic effect is not completely devoid of any practical value.

5 1984: Peirce in the labyrinth

In Semiotica e filosofia del linguaggio (1984a) [Semiotics and the Philosophy of Language, 1984b] ^[14] Eco adds to the theory of interpretation by improving the notion of encyclopedia, “the only model capable of accounting for the complexity of semiosis, both on the theoretical level and as a regulative hypothesis on actual processes of interpretation” (1984a: 255, my translation). ^[15] Eco explicitly criticizes the notion of code.

It is thought that the model of encyclopedia is the most important of Eco’s contributions to semiotics. It has been said that the theory is not complete and that it lacks the rigor of other semantic theories. ^[16] I believe that this contributes to its value, and is not a flaw. Theories that open the way to further inquiry can be as good as those which organize the data in a reasoned and closed order. The rigorous debate about the best formal semantic model, so lively until the 1980s, has silently turned to more practical, almost humble research, which is channeled into computer science and AI applications. Neurological research, on the other hand, has shown that we are still far from mapping the interpretive processes in the brain at cellular level. Thus, instead of looking for what meaning “really” is in humans, AI researchers try to associate input to output expressions in the best way they can, often relying on learning algorithms, like neural networks, that build rules from the actual use of language or other systems of signs. Research into human interpretation is therefore, once again distant from formal models of meaning.

The encyclopedic model still provides two important positive contributions. First, it gives a good description of cultural influences in text comprehension, and this answers to a need that the globalized world expresses with increasing urgency. Immanent theories of meaning, as in generative semiotics and all biplanar or code-based approaches, provide no explanation of cultural divergences of interpretation if not that the right code has not been used. The question becomes then: “What is the right code?,” but such a question does not necessarily have an answer. Just think of texts like the Bible or the Koran.

And second, the encyclopedia puts together processual and systematic approaches. The task is still in progress, but there is no doubt that this is the right solution. It is widely accepted that interpretation can (and must) be seen at the same time as a system or structure, and as a process or development. Yet, to build a model which is both dynamic and systematic is not an easy job, and surely there is a lot of further research to be done. ^[17] I will expand a little more on this point, and I will explain where Eco took his inspiration from, and how he described his model.

Encyclopedia is the model for an interpretant-based semantics. In this approach interpretation definitely leaves behind any lexicalist restriction, “… the sign is always what opens myself to something other. There is no interpretant that, in equating the sign it interprets, does not expand, even slightly, the sign’s borders” (1984a: 52, my translation). ^[18]

The encyclopedic model represents a goal in Eco’s research. It is the framework of semiotic habits that the infinite colony of interpreters deposit as the result of its continuous acts of signification, like the skeleton created by a colony of coral. It constitutes the point of real synthesis between the dynamic conception of Peirce and the synchronic structuralist view, a true synthesis because it is a new conceptual object in which there are no visible stitches but only various references to the history of philosophy. The aim is to build a non-rigid model of the semiotic competence of individuals, groups and cultures. The model is composed both of empirical data and formal principles, since it presumes that

1) properties are not semantic primitives but interpretants, that is, different expressions that in their turn may become subjects of a further representation; 2) the contexts and circumstances are not infinite but are those which, statistically, according to a hypothesis of standard competence (or with regard to the competence required by a certain co-text), are believed to be part of the encyclopedic competence of the sender or of the receiver.

(Eco 1984a: 119)

The encyclopedic model provides a further limit to interpretation:

A textual semiotics also studies the rules by which the interpreter of a text, on the basis of “signals” contained in the text (or even on the basis of previous knowledge) decides which format of encyclopedic competence is required to deal with that text. In this way also the discriminant between interpretation of a text and the indiscriminate use of it can be established. (1984a: 110, my translation) ^[19]

We may call this criterion the “criterion of implicit competence,” which is the same as Eco’s notion of Model Reader (1979b: 7), seen as the set of instructions that actualize the text in the most coherent way.

6 Rhizome and encyclopedia

Eco compares encyclopedia to the model described by Deleuze and Guattari (1976), called rhizome. Deleuze and Guattari do not explain in their text what a rhizome is. They talk about a “system,” “the book” in the general sense. They seem to privilege the pure formal character, the capacity of the rhizome to be a general model, but with no particular object. A model of discourse, a model of text, a model of the world.

As we have seen, in A Theory of Semiotics (1976a) the encyclopedia was a semantic representation that corresponded to the componential model. In Semiotics and the Philosophy of Language (1984a), Eco seems to use the term encyclopedia to refer to some intensional semantic models, similar to the Model Q, and actually the two texts present many similarities. Above all, the encyclopedic model is said to permit the description of the process of unlimited semiosis and linguistic creativity.

However, in Semiotics and the Philosophy of Language the role of encyclopedia is extended, as we will see, to represent the whole “universe of semiosis” in its form, its contents and its functioning. Deleuze and Guattari’s purpose was slightly different from Eco’s. Nevertheless, Eco in his work follows the French authors. Both texts refer to Rosenstiehl and Petitot’s mathematical study of labyrinths (Rosenstiehl and Petitot 1974; Rosenstiehl 1971, Rosenstiehl 1979, Rosenstiehl 1980; Petitot 1977), but the original rhizomatic theory derives from the political criticism of psychoanalysis and linguistics. ^[20] Deleuze and Guattari explicitly attack the rigid and incomplete models of those disciplines, while Eco keeps his theory separate from ideological discussion. Furthermore, by calling his model “encyclopedia,” Eco calls into play an explicit parallel with the Encyclopédie and the system of culture as conceived by the Enlightenment. Yet, Eco, too, strongly criticizes the traditional semantic models, reducing them to particular cases of a more general theory.

Deleuze and Guattari and Eco divided their texts in numbered points, trying to make a list of the properties of their models. I will follow this scheme.

7 1990: “Please stop the interpretation: I want to step out!”

In the 1980s the approach to text analysis called deconstruction, and its inventor, Jacques Derrida, were very popular, particularly in American universities. Peirce’s unlimited semiosis was often cited as similar to deconstruction, which is a very free way of talking about texts, denying any criteria to validate interpretation.

Eco thus dedicated a book, published in 1990, to exploring the rules of good interpretation. ^[24] In it he completes his semiotics by adding to the theory of sign production of A Theory of Semiotics (how man produces signs), to the theory of text in Lector in fabula and The Role of the Reader (how signs have meaning), and some elements of a theory of reception (how man understands signs) in The Limits of Interpretation.

With regard to the limitative principles mentioned above, Eco’s position is the same as before, but the criterion of internal coherence of the text assumes a more important role (1990: 33). This criterion is a testing ground for the universe of discourse activated by the interpreter, and is framed in the hypothetical-experimental process supported by abduction (see Peirce CP 7.164–255). The element of novelty is, however, the adoption of the Peircean idea of community as the depositary and builder of the image of reality, “The thoughts or opinions that define reality must therefore be linked to a community of experts, and this community must be structured and governed in a manner that takes into account supra-individual principles” (1990: 335, my translation; see Peirce CP 5.264–317, 5.358–387, 5.388–410). ^[25] The instrument of this educational activity is interpretation:

From the moment a community agrees on a given interpretation, it creates a meaning that, if not objective, is at least intersubjective, and it is anyway privileged with respect to any other interpretation obtained without the community’s agreement. The result of the process of universal research moves in the direction of a core of common ideas. (Eco 1990: 336, my translation) ^[26]

The question of the limits of interpretation in Eco is thus resolved by five criteria:

1. pragmatic: the interpretations are based on habit-making;

2. universe of discourse: every interpretation must be placed in a universe of discourse which is part of a theoretically unlimited Semantic Universe (Encyclopedia);

3. implicit competence: every text must be interpreted by means of the competence that actualizes it in the most coherent way with respect to the conditions of its production;

4. internal coherence: an interpretive hypothesis that renders coherent a part of the text must also work on the entire text;

5. communitary: the tendency of a community of interpreters of a text to agree on common hypotheses.

Taken together, these criteria constitute the starting point for the foundation of a pragmatics of textual inquiry (taking text in its wider sense) built by extending the theory of interpretation of Charles S. Peirce.

8 Conclusion

Umberto Eco took the basic concepts of his semiotics from Charles S. Peirce, and after three decades, also adopted Peirce’s founding approach. Peirce, as the study of his papers and manuscripts continues to develop, appears more incontrovertibly as an outstanding figure in Western thought. Thus, Eco’s adoption is the acknowledgment of an inquiry of rare determination and depth. Yet, the accuracy of Eco’s analysis, and the way he expands Peirce’s theory and throws light into its darkest corners, renders his work an original continuation of the path opened by the American philosopher. At the same time, Eco grafts onto the tree of interpretive theory the most suitable tools of the European structuralist tradition and the Italian aesthetic school. And, last but not least, he dresses even the most arid topic with his personal wit.