Skip to navigation – Site map
Symposia. Eco and Pragmatism

Duck or Rabbit? Umberto Eco’s Structural Pragmatics

Valentina Pisanty


In this paper I will discuss the extent to which Umberto Eco’s Semiotics maintains the unstable and oscillatory equilibrium between conflicting matrices that is proper to humorous thinking and ambiguous figures such as the famous duck-rabbit illusion and Penrose’s impossible trident. To do so I shall summon the concept of bisociation (Koestler 1964): though never an item of Eco’s own philosophical toolbox, bisociation plays an important role in the creation of some of Eco’s most innovative theoretical contributions, insofar as they result from a possible/impossible graft between Structuralism and Pragmatism. For each, the question will be whether the implant produces self-consistent theoretical amalgams, or whether some degree of contradiction inevitably remains to generate the abovementioned duck-rabbit effect. I shall then analyse the role that such residual incongruities play in a theory that thrives on the self-voiding logic of humorous short-circuits.

Top of page

Full text

1. Bisociation

1You either see the duck or the rabbit. Attention flickers from one perceptual matrix to the other, without any possible synthesis between the two, in a way that resembles the incongruous effects produced by a verbal double entendre or indeed any other kind of humorous device. The point I shall be discussing in this paper is whether and to what extent Umberto Eco’s Semiotics – a lifelong attempt to combine “the devil and holy water,” as he once jokingly defined Structuralism and Pragmatism – maintains the same unstable and oscillatory equilibrium between conflicting matrices that is proper to ambiguous figures and humorous thinking. To do so I shall summon a concept that, though never an item of Eco’s own philosophical toolbox, provides a suitable key to access his peculiar patterns of thought.

Bisociation: “the perceiving of a situation or idea, L, in two self-consistent but habitually incompatible frames of reference, M1 and M2. The event L, in which the two intersect, is made to vibrate simultaneously on two different wavelengths, as it were. While this unusual situation lasts, L is not merely linked to one associative context, but bisociated with two.” (Koestler 1964: 35; original emphasis)

Figure 1: Bisociation (Koestler 1964: 35)

2In his study on The Act of Creation, Arthur Koestler famously analysed the common mechanism underlying three mental activities that, according to him, epitomize the heights of creative thought: Humour, Scientific Discovery and Poetic Invention. In all three the spark of creativity is generated by the surprising encounter of two “self-consistent but habitually incompatible frames of reference, M1 and M2” that intersect in a given point (L). In the rabbit-duck example, L coincides with the graphic marks that may be interpreted both as rabbit attributes (say the ears) and as duck attributes (the beak) without, however, being interpretable simultaneously as both. In scientific discovery, the point of junction corresponds with a familiar element (say the branch of a tree that a chimp unthinkingly chews on) that may be reframed in terms of a different pattern of behaviour (ripped from the tree, it could become an instrument with which to reach a distant banana). In poetic invention L is what some semiologists would call an “isotopic connector,” for example a polysemic word that may trigger a variety of semantic frames.

  • 1 “When two independent matrices of perception or reasoning interact with each other the result […] i (...)

3Although each bisociated matrix is per se compatible with L, attempts at condensing them in a single self-consistent super-matrix are hindered by the habit of considering them as reciprocally unrelated. Yet the interpreter itches for some form of amalgam or unitary composition: the elegant way M1 and M2 both fit in L suggests that it might be possible to do so. Bisociation lasts for as long as the synthesis is prefigured but not yet reached. Like Tzvetan Todorov’s effet fantastique (Todorov 1970) its interval is that of a state of hesitation. What determines the difference between Scientific Discovery and Humour (the case of Poetic Invention is more difficult to pin down as it shares traits of both domains) is whether or not it is indeed possible to retrieve or construct a consistent juncture of the apparently unrelated frames: Eureka! and laughter are alternative outcomes of the same interpretive conundrum.1

4The manifold ways in which the concept of bisociation (though never referred to as such) appears in Eco’s academic and literary corpus is so striking that we could safely consider it as one of the most prominent features of his mode of thinking, as well as one of his favourite objects of study. To start with, his texts are ridden with examples of word-play, mots d’esprit, nonsense rhymes, self-voiding texts, optical illusions, ambiguous and impossible figures such as Penrose’s devil’s fork, analysed by Merrell (1981), that Eco (1990) uses to exemplify the concept of impossible possible worlds: a perfect case of clash between incompatible matrices. Not to mention his passion for duck-billed platypuses, unicorns, hypatias and other hybrid creatures that populate his imaginary landscape. Creative metaphors, portmanteau words and mixed idioms belong to more or less the same realm of open possibilities, serendipitous encounters and unexpected short-circuits between distant frames. His obvious attraction for perceptual, linguistic and intellectual juxtapositions is also testified by the attention he devotes to Abduction as a key-concept of interpretive Semiotics, where – following Peirce – he puts the emphasis on the interpreter’s ability to bring together hitherto unrelated segments (the farther the better) of the Global Encyclopaedia. An ability that Eco himself applies to his numerous analyses of various communicative artefacts, from avant-garde art to popular literature, where the gist of his method often coincides with an unexpected encounter of conceptual frames (hence his famed capacity to mingle high- and lowbrow culture in all the domains of his multi-faceted activity). Finally, bisociation plays an important role in the creation of some of Eco’s most innovative theoretical contributions – starting from the concept of Encyclopaedia, possibly at the very core of his semiotic theory – insofar as they almost invariably result from a graft between separate and sometimes conflicting philosophical paradigms.

5In the next few paragraphs I shall examine some of these implants, in a bid to show how, in a Structuralist framework such as the one that hegemonized the semiotic scene in the Sixties and Seventies, Pragmatism plays the part of the second, colliding matrix that sparks off Eco’s bisociations. For each, the question will be whether the clash produces self-consistent theoretical amalgams, or whether some degree of contradiction inevitably remains to generate the abovementioned duck-rabbit effect. I shall then analyse the heuristic role that such (for the time being hypothetical) residual incongruities play in a Semiotic theory that thrives on the self-voiding logic of humorous short-circuits.

2. Sign Functions and Semiosis

6Judging from the bibliographical references in his published essays, Eco’s first encounter with Peirce dates back to sometime around the late 1960s/early 1970s, between La struttura assente (1968), where Peirce’s name does not yet appear, and Segno (1973), which includes the Collected Papers in the bibliography. Possibly suggested by Roman Jakobson’s 1953 puzzling definition of Peirce as “one of the greatest pioneers of structuralist linguistic analysis” (now in Jakobson 1963 [1992: 7]), the idea that Peirce’s Pragmati(ci)sm may be considered as the other face of Saussurean Linguistics plays a crucial role in Eco’s attempt at formulating “a Unified Semiotic Theory of Signs” (Eco 1973: 20) suitable for the study of both linguistic and nonlinguistic (iconic, indexical, natural…) signs and sign systems.

7The intersecting point (L) between the two self-sufficient Semiotic paradigms is clearly the very term Sign, or semeion, which both Peirce and Saussure choose as the defining element of their new disciplinary fields. But the problems arise as soon as the common link is intentionally and extensionally defined: though somehow similar, the Peircean and Saussurean concepts of Sign are by no means identical, nor do the phenomena they refer to overlap completely. The first is triadic, inferential, mental and exceedingly inclusive, while the second is dyadic, conventional, rigorously linguistic and of very limited extension if compared to its pragmatist counterpart. How to integrate the two perspectives – what to select in each, what to leave out, and how to translate one in terms of the other – is the first scientific challenge that Eco, along with other semiologists who agree with his comprehensive approach, is faced with.

  • 2 “I propose to define as a sign everything that, on the grounds of a previously established social c (...)
  • 3 “A general semiotic theory will be considered powerful according to its capacity for offering an ap (...)

8Right from Segno (and through A Theory in Semiotics, The Role of the Reader, Semiotics and Philosophy of Language, The Limits of Interpretation, down to Kant and the Platypus) Eco’s tendency is to consider Peirce’s definitions as the most satisfactory descriptions of the semiotic device, in line with a longstanding philosophical tradition that dates back to Aristotle, the Stoics and St. Augustin, as well as with the everyday uses of the word (Sign as symptom, clue, footprint, omen…). Hence his attempts to englobe the linguistic sign (roughly Aristotle’s and Peirce’s Symbol) in a wider picture, at first through a series of bold approximations (see diagram below), then by alternatively (i) taming or watering down some of the most irreducible characteristics of Peirce’s Sign/Interpretant/Object triad that prevent it from being trimmed down to Saussure’s Signifiant/Signifié,2 or by (ii) loosening the structuralist prerequisite of a tight pre-existing convention by admitting cases of radical (= codeless) interpretation in the realm of semiotics.3

Figure 2: The Semiotic Triangle (Eco 1973: 26)

9As the clash between structuralism and pragmatism becomes more acute in Eco’s hybrid theory, the notion of Sign is replaced by two different and seemingly opposite concepts. The first, which derives from the ultra-structuralist approach embodied, more than by Saussure’s Cours, by Hjelmslev’s Prolegomena (1943), dissolves the concept of Sign in the myriad of oppositional relations that make up the two planes of language, Expression and Content, in an overall algebraic relation between the elements of the sign-function. From a strictly formal and immanent point of view, the entity we normally call Sign is but “the manifest and recognizable end of a net of aggregations and disintegrations constantly open to further combinations. The linguistic sign is not a unit of the system of signification; it is, rather, a detectable unit in the process of communication.” (Eco 1986: 21). And if – as Hjemslev would have us believe – any process of communication presupposes a system of signification (a code or a language), the sum total of whose possible combinations is virtually included in the folds of its structure, it follows that the contingent and provisional manifestations of such an abstract system become supernumerary entities within a rigorous theory that aims at explaining how any (existing or possible) language works.

10At the same time, Eco follows Peirce’s lead by absorbing the static notion of Sign in the incessant flow of semiosis that constitutes the very essence of thought in action. According to the criterion of interpretability at the core of the pragmatist paradigm, “a sign is only a sign in actu by virtue of its receiving an interpretation, that is, by virtue of its determining another sign of the same object. This is as true of mental judgments as it is of external signs.” (CP 5.569; original emphasis). And since an interpretation – be it mental or externally expressed – is itself a sign in progress, it will in turn trigger another interpretation (an uttered sentence, a silent thought, or indeed any other internal or external representation referring to the same Object as related to the original Sign), and so on, in a potentially endless string of Sign/Interpretations (= Interpretants) whose sum total coincides with the totality of an individual’s mental existence (“the symphony of our intellectual life,” CP 5.397).

11As a cognitive element, the Peircean Sign is therefore the node or intersection point of a bundle of synapses – interpretive habits or Interpretants – that link up to other nodes, and so forth, in an open-ended process that may remind us of a (more or less disciplined) stream of consciousness, while maintaining a connection with the Object (“something else”) that sparked off the previous links in the chain of Interpretants. Sometimes these Interpretants assume the form of perceivable behaviours (external signs such as somatic expressions, gestures or linguistic utterances) which may in turn become the focus of other interpreters’ semiosic activities. Although Peirce is not particularly interested in the communicative aspects of semiosis, within his framework there is nothing to prevent an individual from expressing the results of her musements and inquiries through external signs, thus sharing them with other members of the interpretive community. Culture begins when the Interpretants are made public.

  • 4 Cf. Eco 1983 and 1986 on undercoded and overcoded abductions.
  • 5 See Eco 1976 on the different modes of Sign-production.

12Here is where Eco’s bisociations light up. Once a certain connection between Signs is registered and made public, it acquires a cultural status that differentiates it from all other private chains of Interpretants. If the shared connections prove to be collectively advantageous under some respect or capacity – if they contribute to making the world a more intelligible or predictable place, if they enable people to perform certain tasks, if they promote social cohesion, etc. – they become more and more deeply engrained in the interpretive habits of that specific community. This is how an originally audacious hypothesis may undergo a process of standardization that little by little turns it into a semi-automatic interpretive habit shared by different members of the community.4 It is also how a given behaviour (say an angry expression combined with a raised fist) may lose its original iconicity and become progressively stylized, insofar as the interpretations it triggers depend less and less on the inferential acumen of the single receivers and more on the overcoded Interpretants that the Sign is publicly associated with.5

13By focusing entirely on this category of public Interpretants, Eco brings his Semiotic research back on the culturalistic tracks that were proper to Structural Linguistics. Whatever happens in the mind of a single interpreter, the only aspects of her cognitive activity that are semiotically relevant are those that could be made accessible to other members of the cultural community, or even to members of other communities, provided some form of cultural exchange between the two is underway. In any case, what is left out of the semiotic field are the idiosyncratic connections that only the single interpreter would be able to activate on the exclusive grounds of her private experiences (unless, of course, such experiences were made public through the production of external signs).

  • 6 Whether the words of a language are the outcome of a process of standardization or of a stipulated (...)
  • 7 See Eco 1976 on contextual and circumstantial selections.

14In this perspective, Saussure’s arbitrary relation between the two faces of a sign-function could be reframed in terms of an overcoded interpretation whose cultural crystallization has invested the relation between the Interpretant and its Sign with the status of an almost biconditional rule (not just “if fire, then smoke” but also “if smoke, then [most probably] fire”). The cumulative effect of this over-codification is a relation of near equivalence (fire ≅ smoke) whose application requires only a minimal inferential effort on the interpreter’s part. In other words, near equivalence is a specific form of semi-mechanical, almost compulsory inference. And even in the relatively few cases in which the equivalence is established a priori, through a convention stricto sensu rather than a progressive sedimentation of successful interpretive habits6 (i.e., when a linguistic institution decrees, by means of a metasemiotic instruction, “from now on let x ≡ y”; for example let three blows of the whistleend of match), the application of such a conventional rule is never totally automatic inasmuch as it is context-sensitive and needs to be fine-tuned to the specific circumstances in which the rule is summoned.7

3. Dictionary and Encyclopaedia

  • 8 “Inferential processes (mainly under the form of Peircean abduction) stand at the basis of every se (...)
  • 9 On the combinatory logic of cultural innovation see also Eco 2004.

15Having agreed to the inferential (triadic) nature of all semiotic activity,8 Eco goes back to the structuralist idea that a system of signification virtually entails all its possible communicative outcomes. A strong advocate of the combinatory logic of cultural production (“Let no one say that I have said nothing new: the arrangement of the subject is new,” Eco 1976),9 he is fascinated by all historical attempts – no matter how abortive – to put into practice the principles underlying the Lullian Ars Magna (1305). Just like Lully’s rotating discs produced hundreds of permutations on the basis of a limited number of combinatory units, other more or less utopic efforts to crack the ultimate codes of human sign- and sense-production are based on the search of a handful of primary elements whose possible arrangements or derivations produce a vertiginous quantity of complex meanings. Leaving aside the gallery of heroic failures collected by Eco in his 1997 monograph on the search of the perfect language, the attempts he is most interested in are Dictionary-based semantic models such as Porphyry’s tree, Katz and Fodor’s generative semantics (1963) and – especially – Hjelmslev’s glossematics (1943).

16Dictionary semantic models share the common belief that the incalculable number of existing and/or conceivable sentence- and word-meanings in any given language may be boiled down to a closed set of defining traits. Be they universal categories, innate ideas, essential properties, semantic markers or figurae, the basic quality of such allegedly primary elements is their capacity to generate more complex linguistic or conceptual contents by way of sheer combination. No surplus of meaning is attributable to the accidental information that comes from experiences of the outside world (you don’t need to have seen a cow to grasp the linguistic meaning of the word “cow”): a set of self-sufficient recursive rules applied to the primary elements of a closed system allows for the generation of a multitude of Content Units. Which means that all possible Content Units are virtually contained in the immanent grammar or matrix that generates them: the Code of all codes, no less.

17In accordance with this line of reasoning Hjelmslev believes that, within the structure of any given language, it is possible to “analyse the entities that enter the unrestricted inventories purely into entities that enter the restricted inventories” (Hjelmslev 1943: 71). Granted that for every natural language the huge variety of articulated expressions may be reduced to a manageable repertoire of twenty or thirty phonemes whose possible combinations unleash the immense expressive power of that language, it should be possible to apply the same recursively analytical procedure to the other linguistic plane in order to identify the ultimate building blocks of all possible meanings. The advantage offered by Dictionary models is that they suggest “a sort of molecular landscape in which what we are accustomed to recognize as everyday forms turn out to be the result of transitory chemical aggregations and so-called ‘things’ are only the surface appearance assumed by an underlying network of more elementary units” (Eco 1976: 49): a perfect realization of the wildest Structuralist dream. The disadvantage is that they don’t work.

18Eco (1976 and 1986) spends much time and energy demonstrating their logical and epistemological untenability. More than strictly necessary, as a matter of fact, considering the immediately persuasive arguments with which it is possible to prove that there are no semantic equivalents of phonemes, that the two planes of language are not isomorphous, and that even if “female bovine” were a satisfactory definition of “cow” (which it is not), there would be no reason to believe that “female” + “bovine” belong to a more restricted inventory than the one hosting the allegedly more complex Content Unit “cow.” Yet Eco delves deep into the combinatory logic of Dictionary semantics before dismissing it as defective. He explains the details of each theory with scrupulous attention, as if complying with its method, absorbing its rationale and hoping that the system will work after all. Only when the theory betrays its most evident flaws does he decree its failure to fulfil the expectations it created. The disproportionate efforts Eco devotes to his Anti-Porphyry, Anti-Katz and Fodor and Anti-Hjelmslev polemics suggest that these orderly models exert an intellectual appeal on him that goes beyond the mere mechanism of refutation.

19Neither is it enough to attribute Eco’s passion for defectively airtight systems to his thomistic background, on the one hand, or to his acquired distrust towards all forms of “strong thought,” on the other. Beyond or beneath the biographical explanations lies the author’s typically bisociated pattern of thinking. In a well-told joke, the punchline comes after a longwinded preparatory phase in which the listener is made to feel comfortable with the assumptions of an apparently self-consistent (and often self-assertive) frame. Likewise, Eco builds up familiarity and expectations regarding Porphyry’s Tree and its derived models so that, when the reader is finally caught up in the paradigm’s logic, the collision with an incompatible counter-logic may release the humorous effects of an overwhelming contradiction.

20What, then, is the second frame whose sudden impact opens a breach in the sealed logic of Dictionary semantics? Whenever Eco wants to explode the pretentions of an all-too-orderly sense structure, he confronts it with the worldview of a candid layperson who relies on sensible experience as the ultimate source of truth-judgements. A good example may be found in Kant and the Platypus (“Tiny Tim’s Story,” Eco 1999: 186-91), where a five-year old’s description of watersomething that you put your hands under when you wash them with soap, that is wet and has no shape but goes around all over the place, that is in rivers and bathtubs, etc. – pulls the rug under the analytical definition of water as transparent liquid. The child’s radical empiricism collides with the adult’s rationalist approach. As if puncturing an overinflated balloon, it voids the dream of a perfect taxonomy of its ontological raison d’être. If the way we normally think resembles Tiny Tim’s (who associates each word with an open list of first-hand experiences), then the grandiose masterplan of splitting linguistic meanings down to their primary elements (and then sorting them in a tidy treelike structure) suffers from a cultural fallacy that we may legitimately laugh about, possibly because it says more about our psychological need for closure and order than about the way things really are and work. Closed, organized systems vs. Open, chaotic experiences: this is the first bisociation that Eco calls into play when introducing the concept of Encyclopaedia.

21The next step – which differentiates a purely humorous device from an act of scientific discovery – consists in trying to merge the two divergent frames into one. The link between Openness and Closeness is provided by the concept of Interpretant. Like Koestler’s resourceful chimp, Eco rips the Interpretant – as it were – from the Peircean jungle and grafts it onto structuralist componential analysis, in a bid to infuse life in the sterile denotative and connotative markers through which Dictionary models fail to break down the meaning of words in their simpler components. Insofar as they always open up the Sign to something new, Interpretants introduce an element of creative happenstance in the course of semiosis, while recouping a vital connection between Signs and the outside world. At the same time, however, they may be thought of as relatively stable nodes of (albeit feebly) structured systems, considering the overcoded nature of the semiotic habits they tend to embody. In other words, Tiny Tim’s definitions of water are, yes, idiosyncratic, unstable and disorganized, as you would expect from a child whose knowledge of the world is still sufficiently disentangled from the conditioning of social institutions. But the adult’s remarks and questions – Maybe it’s that red thing in the stove that burns? Is it damp like fruit? So it’s not a solid thing like bread… And if it’s not solid, what is it? etc. – nudge his private Interpretants towards culturally beaten paths, so that little by little he gets used to opposing water to fire, dampness to wetness, liquids to solids, and so on. By conforming to other people’s expectations Tim will eventually become a fully-fledged member of his linguistic and interpretive (= cultural) community.

22Being a member of a cultural community means knowing how to trigger more or less the same Interpretants as most other members presumably would in the same given circumstances. A standardised network of interpretive habits – Eco (2014) refers to it as the Average Encyclopaedia – guarantees a certain degree of automaticity in communicative exchanges, with the obvious adaptive advantage of minimizing misunderstandings within the group. And since the impulse to categorize seems to be part and parcel of human nature, the common Interpretants tend to cluster in the hierarchic and differential patterns that are dear to Structuralist semantics: havens of (local and provisional) Order carved out of the garble of the Global Encyclopaedia.

4. Immanence and Interpretability

23By grafting the Peircean Interpretant onto structuralist componential analysis Eco tries, as it were, to introduce a Carbon-based molecule in the inorganic chemistry of sememes. His revised semantic model (1976: 105-20) overcomes some of the aporias of componential analysis through the introduction of contextual and circumstantial selections that register the Interpretants most frequently associated to each specific sign, be it a word, an image, a gesture, or any other. Given the indefinitely open array of effects that a word such as fisherman could produce when inserted in specific communicative contexts, and granted that some contexts are more selective than others, any average English speaker is expected to know that normally this Content Unit is associated to a range of probable discourse situations; namely, discourses having to do with recreational or commercial fishing, angling or trawling, etc., where each contextual selection triggers some of the possible fisherman Interpretants while it narcotizes other less relevant ones. But the choice of which Interpretants are to be considered relevant is not entirely up to the speakers or the listeners of every single utterance. An overall Encyclopaedic network regulates the different sense-paths that may branch off the common sememe they all derive from.

24Hence the idea that a sememe is in itself an embryonic text. You just have to pronounce the word fisherman (or draw a picture of a fisherman, or mime the act of fishing…) to evoke a certain number of situations where, on the basis of previous communicative experiences, members of your semiotic community are most likely to expect such a sign mentioned. Your mental associations may be idiosyncratic and autobiographical, but even the more personal synapses are largely influenced by culturally acquired interpretive habits (I, for instance, am inclined to link the word to Pinocchio’s Green Fisherman as well as to the lyrics of Paul Simon’s Lincoln Duncan, but I don’t expect you to do the same). The point is: are these habits already embedded or condensed in the Sign as such, as in a pre-Big Bang situation? Or would it be more appropriate to represent the expanding sememe as a public repertoire of interpretations that come and go according to the various uses people make of them? What, for example, happens to the sememe fisherman when Jesus calls his apostles to be fishers of men, the brothers Grimm publish the popular tale of the Fisherman and his Greedy Wife and a xixth century English chemist invents a popular brand of candies he calls Fisherman’s Friends? These are all Interpretants of the same sign and, sure, they all exploit some of the Interpretants previously attributed to fisherman to create new semantic amalgams that, in turn, are registered by the Encyclopaedia and may thus affect further interpretations. But how correct is it to consider such creative expansions as part of the sememe’s structural matrix, rather than as something that just happened to it under certain unpredictable circumstances? The difference between the revised semantic model (still rooted in the Structuralist tradition) and the cognitively inspired Q-model lies precisely in this: as Eco opens up to interpretive semiotics, the short-circuits between different nodes of the Global Semantic Space become ever more erratic and the spectrum of possible textual expansions of each sememe seems to explode in all directions.

25Allowing that the Global Encyclopaedia is a semiotic postulate that may never be described in its complex totality Eco does not, however, give up on the structural analysis of at least local portions of the Encyclopaedia. In The Role of the Reader he spells out the concept that his previous essays only hinted at: the meaning of a sign – or a text – resides in the indefinitely expandable sum total of the effects (or Interpretants) the sign produces or may potentially produce in someone, and therefore in the open series of its possible interpretations. This is an explicit reformulation of Peirce’s pragmatic maxim, and it apparently transfers the responsibility of meaning construction from the sign as such to the interpreters’ activities: whatever people do with a word, a sentence, an image, a gesture, etc., the sum total of their sign-driven actions constitutes the meaning of that sign. Yet according to Eco there have to be ways to (a) predict the spectrum of activities that a sign may produce on its interpreters; (b) distinguish between relevant and irrelevant interpretive activities, i.e. between interpretations that deserve to be (and indeed stand a chance to be) included in the sign’s expanding meaning and totally idiosyncratic uses that have little or nothing to do with or to add to that public meaning. That is why in the Italian version of The Role of the Reader (Lector in fabula, 1979a) he makes a distinction between first- and second-generation textual theories. The former deny the possibility of “disambiguating” a sign or a text independently from the effective situation in which it is delivered and/or received and, consequently, they deny the legitimacy of any attempt to study language as a structural system that precedes its discursive actualizations. Instead, second-generation theories try to achieve “a wise amalgamation between the two possibilities [thus establishing] points of contact between the study of language as a structured system [and the study] of speeches or texts as products of a language that was or is currently spoken” (Eco 1979a: 13; my translation).

26As usual, Eco’s puzzle is how to put together pieces taken from different and seemingly incompatible sets. The key he picks to try to unlock the conundrum is the idea that “a sememe is in itself an inchoative text and a text is an expanded sememe.” In Eco’s words, it is “the Cardanic joint that can unite the semiotics of a code to the semiotics of texts and speeches” (Eco 1979b: 49).

27Derived from Peirce, this concept primarily refers to the pragmatist maxim according to which “to ascertain the meaning of an intellectual conception one should consider what practical consequences might result from the truth of that conception – and the sum of these consequences constitute the entire meaning of the conception” (Peirce 1905, CP 5.9). In other words, the many lines of conduct that may conceivably arise from the fact of having grasped an idea make up the meaning of the corresponding sign: see, for example, the meaning of Lithium, which Peirce identifies with all the chemical reactions that scientists may produce in a laboratory in reference to that substance they call Lithium. In this light, the meaning of a term includes the indefinite number of propositions that could legitimately contain such term, along with the indefinite number of synapses the term and its expanded propositions could trigger, and the indefinite number of behavioural responses they could elicit. Not just now, but in any conceivable future: a very hazardous concept, if you consider it from a nominalistic perspective, since it implies that, over and beyond the arbitrary nature of linguistic signs, there lies some unattainable yet efficient regularity of nature that assures the continuity of each chain of Interpretants.

28But Peirce is not the only author Eco is thinking about when he formulates his maxim. The other explicit reference is to Greimas (1973: 174), quoted by Eco right before he brings Peirce into play: “As Greimas asserted, a given semantic unit such as ‘fisherman’ carries a potential narrative program in its sememic structure: ‘the fisherman [intended as a thematic role] holds within it, evidently, all the possibilities of its doing, all that one can expect from it in terms of behaviour’.” (Eco 1979a: 19; my translation).

  • 10 See Eco 1997 on the search for the perfect language.

29Whether Peirce and Greimas are pointing to roughly the same concept is by no means evident. One the one hand there seems to be an unexpected family resemblance between the two (another bisociated link?). Could the common denominator be rooted in a profound logical and/or grammatical intuition that overcomes the obvious division between Peirce’s interpretive realism and Greimas’s methodical constructivism? This is Eco’s bet. The key to virtual sememe expansions is hidden somewhere between Peirce’s “Logic of Relatives” and Tesnières’s “Dependency Grammar” (which inspired Greimas’s “Actantial Structure”), and possibly much further back to Francis Lodwick’s xviith century attempt to break down the (allegedly universal) content of verbs such as to fish to the frames of action they imply.10 What do these disparate theories have in common? The idea that a single semiotic element possesses valencies or empty slots that predispose it to form certain types of conjunctions with other elements; that, according to the types of bonds it is capable of forming with other elements, it may unfold one or more predictable lines or patterns of action: the skeleton of a story or, rather, of a range of potential stories; and that these proto-stories are somehow outlined in the sign itself, in a way analogous to the indefinite number of life forms made possible by a given genetic code. A sort of Aristotelean entelechy drives each sign towards its self-fulfilling actualizations, as though the sign’s interpreters were but the contingent agents of an inbuilt transformative process, “from being in posse or in germ” (Peirce, Century Dictionary, 1889: “Entelechy”) to its actual forms of existence, partial and imperfect as these may be.

30On the other hand, however, it is clear that Greimas and Peirce are considering different aspects of the semiotic phenomenon. In line with the Hjelmslevian principle of Immanence (a sign system may be theorized as a purely linguistic, self-subsistent entity), Greimas refers to a specific, synchronic state of language within whose closed structure the item fisherman (itself a bundle of coordinated semantic traits) may be combined with a great variety – but not all – other lexemes to form a vast array of possible narrative syntagms. There are many things the semantic system allows a fisherman to do, including swimming, drowning, drinking, singing, dancing, belching and sleeping… – but not flying. Not because (fisher)men are empirically unable to fly (a fact that is linguistically irrelevant), but because the system of the English language does not contemplate this combinatory option. According to Greimas, the range of possible textual expansions is pre-determined by the affordances of the lexeme’s narrative grammar. And even when – as in Jesus’ fishers of men – the word is used metaphorically to mean something that is literally incongruous, the figure of speech is made possible by the linguistic compatibility between the deep semantic structure of the metaphorical vehicle and that of its tenor (hunters of men would not have done the job).

31Instead, Peirce’s pragmatic maxim (hinged on the principle of Interpretability) has nothing to do with the grammar of a particular state of language: his conception of meaning is that of a dynamic and ongoing process whose only limits are set by the constrictions of logic and of reality. The strings of Interpretants that make up the meaning of a word may escape in n directions, according to the more or less adventurous paths followed by members of the community of interpreters, and the possible expansions of the concept of fisherman (or fish, or water…) are not established in advance by the rules of a nomothetic institution. Rather, a process of natural (or cultural?) selection picks out those Interpretants whose consequences yield the most beneficial effects, i. e. the ones that are most liable to turn the world into a more predictable or liveable or manageable environment, i.e. (from Peirce’s scientific perspective) the ones that in the long run prove to be truer, or at least less false than others.

5. Possible and Impossible

  • 11 On the ambiguities of the concept of intentio operis, see Pisanty 2015.

32What does Eco have in mind when he says that “a sememe is a virtual/inchoative text”? That sense is tightly packed within the narrative, discursive and grammatical structures of a coded unit (such as Greimas’s thematic role: the fisherman as a matrix of possible stories), regardless of whether the sense-paths the code virtually contains are effectively actualized in the course of semiosis? Or that semiosis is born free and then restrained by the Noes of experience which determine the sum of “practical consequences [that] might result from the truth of that conception”? It is impossible to decide. All the more so if we extend the scale of Eco’s principle from the textual expansions of a single sememe (the number of sentences that could contain the term fisherman) to the possible interpretations of a single text (the number of ways one could read Moby Dick). Sometimes Eco leans towards the structuralist option, as when – in order to defend the “rights of the text” against the uncontrolled intentions of its deconstructive readers (Eco 1990) – he outlines the concept of intentio operis (that which the text “intends” to say) as the structural matrix capable of generating all possible interpretations. Other times he favours the pragmatist approach, as when he defines the same intentio operis in negative terms (much like Peirce describes our inferential relation with reality) as something that manifests itself indirectly whenever a particular reading collides with the literal evidence which may be drawn from the text.11

  • 12 Regarding this point, cf. Violi 2017: “The very idea of the sememe as a virtual text is quite close (...)

33Both solutions are per se insufficient for the purposes of the encyclopaedic-inferential theory pursued by Eco. Greimas’s thematic role is not designed to explain what happens when someone interprets a text.12 But Peirce’s pragmatic maxim does not describe the culturally determined structures within and against which most communicative acts are performed. Yet neither aspect may be overlooked by a theory that aims at accounting for what actually happens when someone tries to make sense of the world (and the texts that inhabit it) through the specific cultural filters that regulate her system of expectations, as well as through the cognitive procedures that make any interpretation possible.

34“How is it possible for the two points of view to coexist?,” asks Eco (1999: 251) almost a quarter of a century after A Theory of Semiotics. Still grappling with the question that set off his semiotic inquiry, the answer he arrives at in Kant and the Platypus is anything but conclusive.

The result of the preceding reflections is that they must coexist, because if we choose one of them only, we cannot account for our way of knowing and expressing what we know. It is indispensable to make them coexist on a theoretical level, because, effectively speaking, on the level of our cognitive experiences we proceed so that we run – if the expression does not seem to reductive – with the hare and hunt with the hounds. The unstable equilibrium of this coexistence is not (theoretically) syncretistic, because it is on the basis of this happily unstable equilibrium that our understanding proceeds. […] In the process of understanding, the structural moment and the interpretative moment alternate and complement each other step by step. (Eco 1999: 251-3; original emphasis)

35Eco’s argument may at first seem circular: the two points of view coexist because they have to (cf. “Why is there Being rather than nothing? Because there is,” Eco 1999: 17; original emphasis). Is this a case of what Peirce (CP 5.382-3) would call a priori reasoning? Perhaps. But Peirce also says that when science fails, the a priori method is the next best option, “since it is then the expression of instinct which must be the ultimate cause of belief in all cases” (CP 5.383). In want of a better explanation, Eco follows the instinct telling him that order and adventure (cf. Paolucci 2017), categorization and observation, recognition and invention, and all the other bisociated matrices we have discussed in this article are complementary opposites referring to the two forces that, together, drive all semiotic activity. Rather than considering the duck/rabbit (now hound/hare) effect as an undesirable outcome of his theory, he reframes it as the ultimate object of semiotic investigation. If the theory oscillates from one perspective to the other without reaching a satisfactory synthesis, it is because that “happily unstable equilibrium” is the proprium of human understanding.

  • 13 Regarding the same puzzle, Eco (1999: 427) also writes: “It would seem difficult to handle two imag (...)

36But if bisociation is within semiosis itself, engrained in our way of making sense of things, then any attempt at theorizing it in a scientifically consistent way (rather than maintaining it in a state of humorous suspension) meets the obvious hitch of being itself a semiotic activity, based on the same self-contradictory logic it is trying to grasp. A bisociated mind trying to make sense of its being bisociated is reminiscent of the paradoxical situation, described by Eco (1999: 320-1), of a third prosthetic eye attached to the tip of the index finger and pointed towards the other two. What would one see? The mind boggles. “Would we see the index finger with the eyes in our head, or the eyes in our head with the index finger? Once more, either we go by zones of focus (we imagine, alternately closing the eyes in our head and the eye in our finger) or we slip in complete imaginative confusion.”13

Figure 3: Penrose’s Devil Fork/Impossible Trident

37Notice how this hypothetical solution – the alternation between two self-consistent yet reciprocally incompatible visions – is in turn akin to the way Eco (1994: 100 and 1999: 318) defines the concept of possible impossible worlds. With reference to Penrose’s trident (analysed by Merrell 1981), he observes that, whereas the overall two-dimensional figure describes an impossible three-dimensional whole object, each partial 3-D object that may be conjured up by covering one of the two sides of the dotted line is in fact in itself possible. The intermediate area is compatible with both – hence its ambiguity. Like Koestler’s bisociated matrixes, the two possible partial objects become impossible only when merged into one.

38But impossible with respect to what logic? The question will not go away. Maybe such an object could exist in some different world, and it is only because of our intellectual incapacities that – so far – we have been unable to conceive it. The perverse pleasure that we take in analysing the self-voiding logic of such possible impossible worlds is that of our cognitive defeat. A defeat that keeps challenging our scientific imagination: who knows, one day we might be able to “redesign our brain” (Eco 1999: 427) in order to solve the overall puzzle. In the meantime all we can do is to keep trying with the one we have. “Therefore I hold that when we refer to inconceivable entities, we behave as if, on being faced with our ‘white box,’ we were to peep into it by alternately lifting opposite sides of the lid for a few millimetres.” (Eco 1999: 321). Is this a description of what Eco’s structural pragmatics does with the possible impossible object that is semiosis?

Top of page


Eco Umberto, (1968), La struttura assente, Milan, Bompiani.

Eco Umberto, (1973), Segno, Milan, Isedi.

Eco Umberto, (1976), A Theory of Semiotics, Bloomington, Indiana University Press.

Eco Umberto, (1979a), Lector in fabula, Milan, Bompiani.

Eco Umberto, (1979b), The Role of the Reader: Explorations in the Semiotics of Texts, Bloomington, Indiana University Press.

Eco Umberto, (1983), “Horns, Hooves and Shoes,” in Umberto Eco & Thomas A. Sebeok (eds), The Sign of Three: Dupin, Holmes, Peirce, Bloomington, Indiana University Press.

Eco Umberto, (1986), Semiotics and Philosophy of Language, Bloomington, Indiana University Press.

Eco Umberto, (1990), The Limits of Interpretation, Bloomington, Indiana University Press.

Eco Umberto, (1997), The Search of the Perfect Language, New York, HarperCollins.

Eco Umberto, (1999), Kant and the Platypus: Essays on Language and Cognition, New York/San Diego/London, Harcourt Brace & Company.

Eco Umberto, (2004), “Combinatoria della creatività,” conference held in Florence for the Nobel Foundation, [].

Eco Umberto, (2014), From the Tree to the Labyrinth: Historical Studies on the Sign and Interpretation, Cambridge, Massachusetts/London, Harvard University Press.

Greimas Algirdas J., (1973), “Les Actants, les Acteurs et les Figures,” in Sorin Alexandrescu, Roland Barthes, Claude Bremond, Claude Chabrol, Algirdas J. Greimas, Pierre Maranda, Siegfried J. Schmidt & Teun A. Van Dijk (eds), Sémiotique narrative et textuelle, Paris, Larousse.

Hjelmslev Louis Trolle, (1943), Prolegomena to a Theory of Language, Madison, Wisconsin University Press.

Jakobson Roman, (1963 [1992]), Saggi di linguistica generale, Milan, Feltrinelli.

Katz Jerrold & Jerry Fodor, (1963), “The Structure of Semantic Theory,” Language, 39, 2, 170-210.

Koestler Arthur, (1964), The Act of Creation, New York, Penguin.

Merrell Floyd, (1981), “On Understanding the ‘Logic’ of Understanding: A Reincarnation of Some Peircean Thought,” Ars Semiotica, 4.1, 161-86.

Paolucci Claudio, (2013), Strutturalismo e interpretazione, Milan, Bompiani.

Paolucci Claudio, (2017), Tra ordine e avventura, Milan, Feltrinelli.

Peirce Charles Sanders, (1931-1966), The Collected Papers of Charles S. Peirce, 8 vols., C. Hartshorne, P. Weiss & A. W. Burks (eds), Cambridge, Harvard University Press

Pisanty Valentina, (2015), “From the Model Reader to the Limits of Interpretation,” Semiotica, 206, 37-61.

Pisanty Valentina, (2017a), “Che c’è da ridere? Eco e i Protocolli dei Savi Anziani di Sion,” Il verri, 63, Le letterature di Eco.

Pisanty Valentina, (2017b), “The Sememe as a Virtual Text,” in Bent Sorensen & Torkild Thellefsen (eds), Umberto Eco in His Own Words, Berlin, De Gruyter-Mouton.

Quine Willard van Orman, (1969), “Foreword,” in Lewis David, Convention: A Philosophical Study, Cambridge, Harvard University Press, i-xii.

Todorov Tzvetan, (1970), Introduction à la littérature fantastique, Paris, Éditions du Seuil.

Traini Stefano, (2005), Le due vie della semiotica. Teorie strutturali e interpretative, Milan, Bompiani.

Violi Patrizia, (2001), Meaning and Experience, Bloomington, Indiana University Press.

Violi Patrizia, (2017), “Encyclopedia: Criticality and Actuality,” in Sara G. Beardsworth & Randall E. Auxier (eds), The Philosophy of Umberto Eco, Chicago, The Library of Living Philosophers.

Top of page


1 “When two independent matrices of perception or reasoning interact with each other the result […] is either a collision ending in laughter, or their fusion in a new intellectual synthesis, or their confrontation in an aesthetic experience. The bisociative patterns found in any domain of creative activity are tri-valent: that is to say, the same pair of matrices can produce comic, tragic, or intellectually challenging effects.” (Koestler 1964: 45; original emphasis).

2 “I propose to define as a sign everything that, on the grounds of a previously established social convention, can be taken as something standing for something else.” (Eco 1976: 16; original emphasis).

3 “A general semiotic theory will be considered powerful according to its capacity for offering an appropriate formal definition for every sort of sign-function, whether it has already been described and coded or not. So the typology of modes of sign-production aims at proposing categories able to describe even those as yet uncoded sign-functions conventionally posited in the very moment in which they appear for the first time.” (Eco 1976: 5).

4 Cf. Eco 1983 and 1986 on undercoded and overcoded abductions.

5 See Eco 1976 on the different modes of Sign-production.

6 Whether the words of a language are the outcome of a process of standardization or of a stipulated convention is not a problem that Eco dwells on extensively, in accordance with the 1866 linguistic ban on the discussions regarding the origins of language. However, he does occasionally seem to slant towards the first hypothesis, for example when – in the last page of The Search of the Perfect Language (1997) – he mentions Ibn Hazm’s argument against the idea that different languages were born from convention: “if so, people would have to have had a prior language in which they could agree about conventions.” Notice the similarity between this xth century objection and Quine’s 1969 “vicious regress” argument in his forward to Lewis’ Convention, p. xi. The difference, of course, is that Ibn Hazm believed that historical languages derived from an original language which included all the others.

7 See Eco 1976 on contextual and circumstantial selections.

8 “Inferential processes (mainly under the form of Peircean abduction) stand at the basis of every semiotic phenomenon.” (Eco 1986: 8).

9 On the combinatory logic of cultural innovation see also Eco 2004.

10 See Eco 1997 on the search for the perfect language.

11 On the ambiguities of the concept of intentio operis, see Pisanty 2015.

12 Regarding this point, cf. Violi 2017: “The very idea of the sememe as a virtual text is quite close to Greimas’s thematic roles. But there is a key difference. Simplifying substantially, we could say that Greimas’s theory is a theory about how text is structured, while Eco’s is a theory of knowledge and interpretation. More precisely, Eco’s is a theory about how our general cultural knowledge is organized and how we use it in order to interpret the world and the various texts that inhabit it.”

13 Regarding the same puzzle, Eco (1999: 427) also writes: “It would seem difficult to handle two images at the same time; perhaps it would be necessary to close the two normal eyes when the third is in use, but I’m not sure this would be enough. The most reasonable conclusion is that the innovation would make it necessary to redesign our brain.”

Top of page


Electronic reference

Valentina Pisanty, « Duck or Rabbit? Umberto Eco’s Structural Pragmatics », European Journal of Pragmatism and American Philosophy [Online], X-1 | 2018, Online since 20 July 2018, connection on 20 August 2018. URL : ; DOI : 10.4000/ejpap.1131

Top of page

About the author

Valentina Pisanty

University of Bergame

Top of page


Licence Creative Commons
Author retains copyright and grants the European Journal of Pragmatism and American Philosophy right of first publication with the work simultaneously licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

Top of page
  • Logo Associazone culturale Pragma
  • Logo DOAJ – Directory of Open Access Journal
  • Logo École des Hautes Études en Sciences Sociales – EHESS
  • OpenEdition Journals