Signs,webs,and memories

Introduction

Since the inception of the cultural turn in the social sciences more than 40 years ago, the relationship between semiotics and sociology has been a complicated one. It has been quite fruitful in methodological terms, but the integration at the general level of semiotic theory and sociological theory has often proceeded through catchphrases and misunderstandings rather than through the rigorous construction of a unified perspective.

Structuralism has come and gone, and come back again in a recent resurgence (Lizardo, 2010, 2013). But the days of a close confrontation between the theoretical implications of structuralism and general theory building seem long over (Giddens, 1979; Glucksmann, 1974; Lemert, 1979a, 1979b; Schwartz, 1981). On the more interpretive, pragmatic side there are signs of a renewed interest in Peirce’s theories, especially for what regards the interest on categories and ontologies (Kockelman, 2013), meaning mechanisms (Gross, 2010), theorizing (Swedberg, 2015), and the logic of inference (Tavory and Timmermans, 2014).

This complicated relationship seems all the more ironic the more we acknowledge that generations of scholars have become accustomed to work, to quote Clifford Geertz, with a definition of culture as ‘essentially a semiotic one’ (Geertz, 1973: 5). With this definition, a whole set of tools has become common currency for many practitioners of cultural sociology and social theory: signs, symbols, codes, narratives, binaries, actants (Latour, 2005), sometimes with very little attention to their original ambiguity and their intellectual history as methodological tools. But, largely, the intellectual space at the intersection of semiotics and sociology is still dominated by various strands of post-structuralism (for a recent review see Han, 2014), or by attempts to fuse into a ‘structural hermeneutics’ (Alexander, 2003) the best intuitions of formal and interpretive approaches (Reed, 2012).

It is my contention that there are alternative routes, which can advance our efforts toward integration, and that these seem all the more necessary the more the ‘success’ of semiotics (as a method and as a perspective) has produced some unintended consequences, which take the shape of several types of fallacies in the adoption of a ‘semiotic stance’ (Kockelman, 2005).

In this article, I want to review and assess the work of Italian semiotician Umberto Eco for what it can contribute to the development of a semiotic sociology, an ambitious (and still incomplete) intellectual position that is only partially overlapping with cultural sociology as we have come to know it after the cultural turn. Focusing on Eco and on the key concepts of his semiotic theory brings to the center some alternative ways to analyze meaning and the semiosic mechanisms associated to its emergence and reproduction. On the one hand, Eco’s work can be read as a relevant contribution to systems theory, since it occupies an intermediate position between those of Niklas Luhmann and other semiotic theorists of culture like Juri Lotman. On the other – and this will be the reading that I will support here – the development of Eco’s ‘general semiotics’ (Eco, 1984) can provide important building blocks for a cognitive theory of culture that has at its center the notion of interpretability and the dissolution of the strong relationship between semiotics and natural language.

The theoretical implications coming from these positions can contribute greatly to a new coming of age of cultural sociological theory. In particular, I want to make a case for the relevance of Eco’s work and its usefulness to address some of the pressing dilemmas that culturally-oriented social theory currently faces: a) the degree of autonomy of cultural structures, b) their foundation of a cognitive sociology, and c) the nature of the relations among elements of the cultural system. In my conclusions, I will set some directives for further work on these problems (which, necessarily, needs to move even beyond the core argument that I intend to present here), and the need for a more explicit semiotic solution in terms of the adoption of a ‘general semiotics’ perspective.

Preliminary positions: Semiotics as a logic of culture

An interest for the internal logic of culture is central in Eco’s work, in particular that part which is most explicitly concerned with addressing theoretical problems in semiotics, a focus that started to become more and more explicit with La struttura assente (Eco, 1968), and the works that updated that manifesto of ‘methodological structuralism’ (Eco, 1971, 1973a, 1973b, 1973c). In A Theory of Semiotics (Eco, 1976), however, a much more formal approach to the issue of semiotics as a logic of culture was presented, under the headings of ‘A Theory of Codes’ and ‘A Theory of Sign Production’, an attempt to reach a synthesis between the two main traditions in semiotics, the one originating with Saussure, Jakobson, and Hjelmslev, and the one that considers semiotics under the banner of Peirce’s own idiosyncratic version of American pragmatism.

A Theory of Semiotics contains both the toolkit that Eco refined for the next four decades (the idea of the encyclopedia and the focus on cognitive semantics) and a series of programmatic statements about semiotics as a discipline endowed with the mission of studying culture. This can be read as a maximalist program: semiotics is ‘a substitute for cultural anthropology’ (Eco, 1976: 27), whose aim is to study ‘the whole of culture’ (p. 6), because ‘the whole of culture is signification and communication and that humanity and society exist only when communicative and significative relationships are established’ (p. 22).

At first sight, this might look like an idealist position, but if we look closer, Eco discusses two different positions regarding the relationship between semiotics and culture: a) ‘culture must be studied as a semiotic phenomenon’; and b) ‘all aspects of culture can be studied as contents of a semiotic activity’ (Eco, 1976: 22), meaning that one could look at culture ‘from the semiotic point of view’, in order to identify the ‘semiotic laws’ that make ‘objects, behaviors and relationships’ function socially. In turn, these objects ‘as object[s], already [have their] own sign-function, and therefore a semiotic nature’ (Eco, 1976: 28). From this point of view, thus, every entity can become a semiotic phenomenon, particularly because ‘cultural phenomena can be and must be seen as significant devices’ (p. 28) that are related to ‘the contents of a possible signification’.

It might seem that Eco reproduces some tenets of structuralism, particularly the distinction between a plane of expression and a plane of content. Yet, while the language and some of the arguments developed by Eco in A Theory of Semiotics are quite explicitly structuralist, the core of his theorizing about semantics is not. And, while a significant portion of that book is entitled ‘A Theory of Codes’ (1976: 48–150), the key idea is not one of perfect correspondence between types in the semantic system and types in the syntactic system (1976: 52–3), but one in which codes are not ‘a natural condition of the Global Semantic Universe nor a stable structure underlying the complex of links and branches of every semiotic process’ (Eco, 1976: 126).

This de-structuration of the semantic space rests, however, on a great structuralist intuition, even though it was subsequently domesticated by structuralism. It is present already in Saussure, who wrote of syntagmatic and associative relations as key mechanisms for the definition of meaning. In the Course in General Linguistics (1959), the idea of meaning that Saussure exposed is relational and differential, and those two types of relations describe two different modalities of difference among units in the semiotic system. Associative relations, which are the ones that matter most in the definition of Eco’s view of signification, deal with the possibility of substitution of a signifier that holds a given position in a syntagm (a chain of signifiers) with another signifier. Contrary to the formalized contrastive relations that Jakobson later characterized as ‘paradigmatic’, Saussure’s idea is that these operations of substitution are performed in the mind of the speaker by mental associations, which can involve the two distinct planes of the signifier or of the signified. The end result of these substitutions is a ‘mnemonic chain’ that links together units that are connected by resemblances in form and meaning, while not necessarily requiring a formal specification of the marked and unmarked traits (for a sociological application see Brekhus, 1998) that compose the meaning of a certain syntactic (or semantic) unit. Quite to the contrary, an associative family of terms involves a series of ‘latent associations’ (Thibault, 1997) that exist in the linguistic system as possibilities, and which can be mobilized in the production of concrete acts of speech.

This network model of the semiotic system is only superficially expressed in Saussure’s work. However, its presence also hints to the centrality of some aspects of semiotics that Eco pursued in closer connection to the ‘other’ classic of semiotics – Charles Sanders Peirce: the fact that a semiotic system involves connections among expressions that result in open-ended processes; the fact that this process results in close-knit clusters of terms, resulting in a different density among sectors of the semiotic system; and finally a consideration for the relational nature of the process of semiosis. However, Eco’s reworking of these principles meant also a turn away from his original structuralism towards a more pragmatic vision of semiotics and communication, epitomized by the concepts of the encyclopedia as a metaphor for the organization of content, and of unlimited semiosis as the primary semiotic process. In this regard, his semiotics incorporated both an attention to cognitive semantics, and to ideas that were traditionally addressed by pragmatics.

The encyclopedia and unlimited semiosis

In Eco’s semiotics, the encyclopedia and the unlimited nature of semiosis are inextricably linked. Peirce’s model of semiosis stands at the center of this complex architecture, which combines cognitive semantics and a theory of the cultural system in what has been identified as a ‘slow and respectful convergence’ (Proni, 2015). However, the reconstruction of how precisely Eco combines semantics and a theory of culture needs some analysis of his interpretation of Peirce’s work, especially because in Eco’s understanding of the encyclopedia as a format for culture one notion, that of the interpretant, holds a central position.

Semiosis, for Peirce, has to be understood as an ‘action, or influence, which is, or involves, a cooperation of three subjects, such as a sign, its object, and its interpretant, this tri-relative influence not being in any way resolvable into actions between pairs’ (Peirce, 1931–58: 5.488, hereafter CP). While common sense can provide some idea about how to regard the ‘sign’ and the ‘object’ (and common sense usually misguides those who try to interpret Peirce’s vocabulary), we are left more or less in the dark when it comes to the term ‘interpretant’ – which, for Peirce, is not the agent who interprets signs but a product of the process of semiosis:

A sign or representamen, is something which stands to somebody for something in some respect or capacity. It addresses somebody, that is creates in the mind of that person an equivalent sign, or perhaps a more developed sign. That sign which it creates I call the interpretant of the first sign. The sign stands for something, its object. It stands for that object not in all respects, but in reference to some sort of idea, which I have sometimes called the ground of the representamen. (CP 2.228)

If it were sufficient, there would be no need to investigate the relationship not between the sign/representamen and the ground/immediate object, but the one that exists between sign and interpretant. If we recall Peirce, the interpretant is above all an ‘equivalent sign, or perhaps a more developed sign’ (CP 2.228), which refers to ‘an object to which [the sign] itself refers, in the same way’ (CP 2.300), in the process leading us to know something more about the object and therefore refining the meaning defined by the ground. This process that leads from one sign to the other (its interpretant), with the latter becoming the sign of a new triadic relation, in the end produces an unlimited semiosis which has ‘an absolute object as its limit’ (CP 1.339), in the form of a habit or a final interpretant (CP 4.536), that is a disposition to act upon the world or a valid argument that draws conclusions from all the premises that can be derived from the accumulation of knowledge about the object. Or, in other words, we can know the object only if we go through a series of endless translations of one sign (as an expression), into others (its interpretants, which do not necessarily belong to the same semiotic system) until – logically – all efforts return us to an adequate knowledge of the object (in this case, the dynamic one, not the immediate one, which is a mental representation).

Following Peirce, Eco maintains that this process is potentially endless: ‘the representation of the content takes place only by means of interpretants, in a process of unlimited semiosis’ (Eco, 1984: 68), in which the succession of interpretants creates the nodes of a network (an ontology) as well as the ties, that is, the realized or available paths that link signs to other signs and which are mobilized by semiosic processes. It is in this respect that Eco speaks of the encyclopedia as not only a ‘semiotic postulate’ and a ‘regulative idea’ that describes the organization of knowledge, but as a universe that is ‘structured like a labyrinth’ (Eco, 1984: 83) that has peculiar characteristics: ‘a) it is structured according to a network of interpretants; b) it is virtually infinite because it takes into account multiple interpretations realized by different cultures’ (Eco, 1984: 83).

Starting with A Theory of Semiotics, and with an accentuation in Semiotics and the Philosophy of Language, the concept of encyclopedia was identified as a format for the organization of cultural content. By reworking Peirce’s triadic model of the sign, it was possible to isolate the universe of the semiotic system both from the object and the subject of perception. In terms very similar to Luhmann’s systems theory, the encyclopedia is a space in which the open-ended character of interpretation is limited as long as successive interpretations (of a sign into its interpretants) make the object of interpretation more specific, or, as Elena Esposito maintains, ‘a continuous production of signs or interpretations that specifies and defines the initial operation, as the coordination of autopoiesis defines the meaning of communication’ (Esposito, 2013).

Yet, if all that exists is the endless chains of interpretants and everything is resolved within the universe of semiosis, the notion of an active subject disappears since it is not necessary for the operation of the system. In this sense, semiotics could well be a type of ‘de-centered’ social theory (Lemert, 1979b), in which both subject and object are bracketed out. In A Theory of Semiotics, this analytic exclusion resulted in the peculiar argument of the book, which started with the theory of codes and ended with the theory of communication. Or, as Eco was to put it 20 years later, with the dynamical object as a ‘terminus ad quem of semiosis’ and not as a ‘terminus a quo’ (Eco, 1999: 3): the dynamical object was to be known, if only asymptotically, only through the series of interpretants that specified the immediate object.

We find, thus, two alternative directions for social theory that can spring from Eco’s ideas about the organization of the encyclopedia: one is rooted on the notion of the autonomy of the cultural system, while the other connects the description of the process of semiotics to a cognitive theory of semantics and to the possibility of a theory of agency. Both cognitive semantics and the potential focus on agency, in turn, constitute potentially useful building blocks for the implementation of a semiotic theory of socio-cultural mechanisms. In order to assess the usefulness of this direction one has, however, to turn to Eco’s later works, in which the two streams – the cognitive dimension of meaning, and the relation of this theory to a theory of the cultural system – are fleshed out. In particular, one has to turn their attention to the typologies of content and the encyclopedia that Eco describes in his major later works, namely Kant and the Platypus (1999) and From the Tree to the Labyrinth (2014).

Semiosis, cognitive content, and ontologies

In Kant and the Platypus Eco addresses specifically the processes involved when semiosis is connected to the reality of human experience, and he does this with a cognitive twist:

When we presume a subject that tries to understand what it experiences (and the object – that is to say, the Thing-in-Itself – becomes a terminus a quo) then, even before the formation of the chain of interpretants, there comes into play a process of interpreting the world. (Eco 1999: 4)

The description of this process is centered on three important concepts in Eco’s later work, those of ‘cognitive type’, of ‘nuclear content’, and of ‘molar content’ (Eco, 1999: 130–43). They represent the key concepts of a small cognitive turn that connects the encyclopedia (as an analytically autonomous network of cultural units) to the processes of selection and interpretation, that is with inputs coming from dynamical objects in the guise of terminus a quo (and especially objects in the external world), as well as with different dimensions of content that are mobilized both in recognition of the correspondence of selected traits of the object to portions of content, and in the production of adequate interpretants that are constrained by the semiotic process of selection (and in this sense, this cognitive orientation provides a pragmatic justification to the idea of the limits of interpretation: Eco, 1990: 203–19, 1992).

Cognitive types (CT) are an element of perceptual semiosis (Eco, 1999: 138), involved in the recognition of tokens (p. 130), the identification of referents (p. 139), and the retrieval of objects (pp. 140–1). The CT is a ‘multimedial’ construct (p. 130) which bring together characteristics of objects that are prima facie used in its identification: not only morphological traits, but also sensory data and ‘functional characteristics’ that are made pertinent by individuals in the process of building a template or schema for successive recognition of the object. The CT is, however, a ‘private’ construct that individuals compile out of sensory experience (including the experience of communication among conspecifics). However, out of the traits that are pertinent in an individual’s CT type, communications produce public collective interpretations that homologate the features of the CTs eventually held by individuals. In other words, interpretants of the cognitive type are produced in a process that leads to communicative consensus.

Eco calls this series of interpretants the nuclear content (NC), a set of public representations through which members of a community of interpretation attempt to ‘homologate’ individual cognitive types by producing ‘verifiable’ interpretants that represent ‘the way in which we try intersubjectively to make clear which features go to make up a CT’ (Eco, 1999: 138). From this point of view, the NC does not coincide with the broad competence an individual may (or may not) have of a specific object of reference: it is a form of condensed knowledge that includes what members of a community of interpretation have in common (on this notion of ‘common competence’ see Eco, 1999: 177ff). Or, in other words, the nuclear content is the series of interpretants most likely to be produced in order to identify a token as belonging to a type, for the retrieval of a token of a given type following a set of instructions that are public and shared. In this regard, the nuclear content constitutes a basic set of cultural tools (in the form of instructions and a finite set of interpretants) that allow basic operations of categorization and reference.

However, culture is more complex than even the widest set of NCs would imply. Nuclear contents are the public sediment of/for cognitive types, and play little role beyond those basic operations and communications between conspecifics. They exclude specialized knowledge, and focus more on the commonalities of content that are distributed more or less in an indiscriminate way within a community. Yet, much of the knowledge of the world we possess (and I interpret the term ‘world’ in a very liberal way) resides outside this basic set of interpretants. Eco, indeed, introduces a further notion, that of ‘molar content’ (MC), in order to account for those sectors of our competence that are not usually mobilized in the process of recognition of tokens. The molar content is a form of ‘complex knowledge’ and ‘broadened competence’ that is not ‘indispensable for perceptual recognition’ (Eco, 1999: 141). In this sense, the molar content is a content on which there exists some social consensus (though with ‘some fraying and gray areas’; p. 141), but which is also unequally distributed among members. While, in fact, the nuclear content is exceptionally widespread, Eco argues that the molar content is definitely specific: it is the kind of ‘specialized knowledge’ (Eco, 2014: 73) that is available for members of particular sub-communities and ‘represents portions of sectorial competence’ (Eco, 1999: 142).

The relationship between NC and MC highlights a fracture (in qualitative terms) within the semiotic space, which is not usually covered by many sociological accounts of ‘meaning’. This happens for two reasons: a) they work with a limited notion of the ‘cultural system’; and b) they still translate this notion of the cultural system in largely structuralist terms. The centrality of the concept of interpretant in Eco’s semiotics addresses this second problem, while the complex organization of the encyclopedia (which I am going to discuss in the next section) reworks the notion of the cultural system both from the point of view of semantics, and from the one of pragmatics.

Ontology reloaded: The encyclopedia and cultural memory

In Kant and the Platypus, Eco’s interest was aimed more at the understanding of the relationship between perception, semiosis and cognition, and the notion of molar content was peripheral, in comparison with the centrality of the discussion of cognitive types and nuclear contents. The molar content was defined as broad competence, but also as a key element of the semiotic cultural space: ‘The sum of MCs coincides with the Encyclopedia as a regulative idea and a semiotic postulate’ (Eco, 1999: 142).

As the key concept that guides Eco’s thinking about semantics (Desogus, 2012), one must look at how it is employed in order to overcome sclerotized notions of the cultural system, with a look at its internal organization and at its role in semiosis. As we have seen, semantic representation is dependent on the relations that are established between a cognitive type and a nuclear content. These relations can be inscribed in the process of semiosis insofar as the nuclear content encapsulates a series of interpretants, which are publicly available and which constitute a kind of intersubjective knowledge about a particular cultural unit. By defining the nuclear content as interpretants, Eco subscribes once more to a model of a semantic encyclopedia, and he connects it to cognitive processes and to the semiosis of perception. In turn, the cognitive types established by perception provide a first ground for the production of ‘knowledge of the world’ and for its encyclopedic organization (Eco, 2014: 3).

However, Eco also argued in favor of an analytic distinction between two different formats for the understanding of content, ‘nuclear content’ and ‘molar content’. This distinction returns in the consideration of the internal organization of the encyclopedia, and especially the difference between a ‘median’ encyclopedia and ‘specialized’ ones. Nor does this distinction exhaust all the possibilities, since in From the Tree to the Labyrinth (2014) Eco presents at least four dimensions of the encyclopedia: a) the individual; b) the specialized; c) the median; and d) the maximal encyclopedia.

The individual encyclopedia is more than a collection of the cognitive types mobilized by an individual agent in the process that leads to interpretation. On the contrary, it is a quite different notion, because it includes all the competence (in terms of knowledge of the world and of the knowledge to act upon this world) that is in principle utilizable by the individual in the process of communication, which for Eco requires the generation of signs and interpretants in a semiosic chain. Some of these competences can be highly individual, and some can be – if no communication occurs – completely private. At the same time, the individual encyclopedia cannot be equated to individual memory, because it also organizes aspects of what Jeffrey Olick identifies as ‘collected’ memory, that is, the ‘aggregated individual memories’ of a group, and of collective memory proper (Olick, 1999).

If we speak of the intersubjective dimension of memory, however, we are reaching dimensions of the encyclopedia that transcend individual competence and acquire a more collective dimension. A median encyclopedia, in Eco’s vision, is a relatively restricted repository that organizes common sense knowledge, the ‘contents of a given culture’ (Eco 1999: 73) that are actually accessible (and most of the times concretely accessed) by members of a collectivity: ‘not […] all of its contents are shared by all members of a given culture, but rather that it is shareable’, and retrievable provided that a members possesses not only some idea of the content (in this case, the set of nuclear contents that have been historically selected), but also of the instruction for having access to it (p. 74). Which, in turn, corroborates the idea that participating in encyclopedic knowledge means navigating from actually stored information to latent information that is externally stored (in a variety of cultural artifacts).

There are in fact outer zones that are not part of the median encyclopedia, although they have a highly formal semiotic organization. In Kant and the Platypus, the molar content represented the specialized knowledge one could have about an object. Accordingly, specialized encyclopedias are highly sectorial, and as such very local, networks of interpretants which organize (although in a fluid and contingent way) a kind of competence from which ordinary members of a given community are excluded in practice, although – by training, socialization, or learning – they can incorporate that competence as part of their individual encyclopedias.

In Eco’s words, the relations among these three dimensions of the encyclopedia are those among networks that only partially overlap. The hypothesis is one according to which there is a

kind of solar system (the Maximal Encyclopedia) in which a great many Specialized Encyclopedias describe orbits of various circumferences around a central nucleus (the Median Encyclopedia), but at the center of that nucleus we must also imagine a swarm of Individual Encyclopedias representing in sundry and unforeseeable ways the encyclopedic notions of each individual. (Eco, 2014: 72)

Specialized encyclopedias vary with time and are able to select their relevant material (a typical instance of this process is Parsons’ famous remark in The Structure of Social Action: ‘Who reads Spencer now?’). As such, their aspiration is to reach a degree of economic organization of their content, by means of forgetting what is actually remembered, and to provide at the same time some accepted mechanisms for the retrieval of forgotten information; in this sense, the notion of memory consists less of a series of collective representations and more of a mechanism for storing and generating content (Lotman, 1993). This guarantees that the creation of new paths and the revitalization of forgotten ones is always a possibility within the system of cultural units.

In this regard, specialized encyclopedias seek both stability and paradigmatic status. The situation is rather different for what Eco identifies with the maximal encyclopedia, a notion that is closer to the idea of the encyclopedia that characterized his earlier work (1976, 1984). Based as it is on interpretants and therefore on unlimited semiosis, the maximal encyclopedia is ‘potentially infinite because it is forever in fieri’, and it extends its reach towards the past and the future, retaining traces of knowledge and allowing room for the production of new interpretants. Truth is not a criterion for the inclusion of interpretants, since the encyclopedia records ‘everything that has been claimed in a social context, not only what has been accepted as true but also what has been accepted as imaginary’ (Eco, 1999: 50), provided that it can enter materially the chain of semiosis.

In conformity with other holistic approaches to the semiotic space (like Lotman’s), the concept of encyclopedia is therefore characterized by an internal discontinuity that affects the form and the paths of the available semantic networks. Within the vast and largely virtual space of the maximal encyclopedia, there exist more recognizable, and dense, zones which have achieved some degree of intercommunicative consensus about the links between interpretants. The specialized and the median encyclopedia are certainly instances of this uneven density in the encyclopedic space. These portions of universe (Eco, 1999: 60) produce their own ‘ontologies’, that is, local representations of a sector of the global semiotic space and shared assumptions about the world that guide interpretation and which are in turn produced by the interpretive activity.

The catalogue of concepts that Eco produced represent a different take on the organization and the mechanisms of operation of the ‘cultural system’, certainly different from many of the neo-structuralist assumptions that orient theorizing and analysis in cultural social theory. Where the latter speaks of codes, signifiers, signifieds, and most of all of cultural structures, the former relies on totally different semiotic assumptions that happily dispose with many of these notions: triadic relationships problematize the biplanar connection between expression and content, interpretants open semantics towards pragmatics (Eco, 1991), the networks that constitute the encyclopedia and its local ontologies push to the extreme the traditional concern for the relational nature of meaning, while in the meantime stressing the primacy of processes at the expense of systems.

At the same time, however, Eco’s semiotics is not immediately translatable in terms of a ready-made sociological theory. At best, it provides tools for the assessment of the strengths and the shortcomings of currently existing versions of the role of culture. However, those tools can be extremely successful in strengthening many of the notions that seem common currency in contemporary cultural sociology, and consequently in some sectors of sociological theory. In the next section, I will discuss some issues in contemporary theorizing that could benefit from a shift in our semiotic focus, and with the role Eco’s semiotics can play in making this shift successful.

Discussion (and a sociological coda)

Attacks from this perspective come from several directions that constitute the vanguard of sociological theorizing about culture. Martin (2010) argues that a cognitive theory and a vision of culture as ‘a complex web of meanings’ (2010: 228) are incompatible because of our cognitive limitations (which affect the way we use memory, process environmental data, and assign dimensionality to our experiences of the environment). These cognitive limitations allow only for rather simple cognitive operations and, while ‘culture may be complex, [that] means it is not in our heads’ (Martin, 2010: 240), as a replication of objectified cultural structures, networks, and artifacts. Lizardo takes a step further, questioning the adequacy of theories of culture as internalized through symbolic communication, opting for a procedural theory that postulates the human inability to internalize ‘cultural content or symbolic representations’ (Lizardo, 2012: 71). Lizardo takes therefore the path of non-representationalism, and to a notion of culture as ‘made personal exclusively in the form of dynamic, action-centered, context-sensitive, non-content-bearing, non-propositional, non-symbolic, sensorimotor-competences’ (p. 74). Both positions hit a crucial point: that our sophisticated reconstructions of culture as an external environment of action (Alexander, 1988, 2003) should at best be conceived analytically, without conflating this analytic and largely artificial modeling and individual cognitive processes. In other words, these positions that attempt to build and legitimize cognitive social science try to limit the space of applicability of a semiotic theory of culture, by reminding us that it works in an ‘analytic’ space and as a ‘methodological’ fiction, the reality of the concrete way in which culture works being at the same time more complex and simpler.

Simpler, in terms of the basic cognitive and bodily operations through which culture is procedurally internalized. More complex, because it denies any legitimacy to any kind of homology between culture as objectified and externalized, and culture as internalized. Many of the critiques challenge Geertzian and post-Geertzian accounts of culture as public and inscribed in symbols (Geertz, 1973), together with its often implicit semiotic methodology. In espousing a Wittgensteinian position, Geertz claimed that ‘culture is public because meaning is’ (at least, contractually negotiated); William Sewell maintained that culture is formed in semiotic relationships, though their coherence may be ‘variable, contested, ever-changing, and incomplete’ (Sewell, 2005: 57); Jeffrey Alexander called for an analysis of culture that privileged ‘the codes, narratives, and symbols that create the texture webs of social meaning’, against attempts in which meaning is ‘simply read off from social structure’ or reified (Alexander, 2003). All these attempts to build a rigorous cultural sociology are not well equipped to counter attacks that come from an aggressively cognitivist approach to culture, and some of the reasons lie precisely in the kind of semiotics they adopt, still burdened by precisely what those authors (chiefly Alexander and secondarily Sewell) perceive as the strength of a semiotic approach: its promise to deliver a logically economic model of cultural organization that encompasses simultaneously formal patterns of relations (in their describable and generalizable features) and mechanisms for the generation of pragmatically successful communications.

If this attempt were limited to the investigation of the ‘analytic’ form of cultural autonomy (Kane, 1991), then it would be another form of idealism (McLennan, 2004, Alexander, 2006) or schematism (as is clear in Sewell’s attempt to overcome the ontological bias found in Giddens’ reworking of the notion of structure: Giddens, 1979, Sewell, 1992). Since they tend to conceive of culture as a relatively autonomous dimension of sociological explanation, however, their search for cultural mechanisms is hindered, rather than enhanced, by their more or less structuralist premises: ‘structuralist approaches often leave the causal mechanisms connecting cultural systems to concrete action implicit in their models’ (Norton, 2014: 166). The solution, therefore, would consist in a better specification of (a) the generative mechanism that link ‘codes’ to ‘messages’ (these terms are awful, I know), and (b) a mechanism for the production of those possible ‘messages’ to concrete situations and circumstances. While a structuralist approach to culture can easily provide us with (a) (in terms of semantics of action, generative grammars, narrative programs, etc.), it cannot provide us with (b), unless it violates some fundamental postulates about the ‘virtuality’ of structure. Pragmatics, after being analytically excluded from the picture, would surreptitiously be brought back in.

How can we keep together signification and communication, if we want to make them the building blocks of cultural explanation? And how does pragmatics intersect these two different aspects of semiotics? I think that Eco’s theory, as I have presented it, creates an opportunity to avoid the many fallacies that one can find in neo-structural analysis, while at the same time not committing itself to some form of reductionism. It does so because it adapts Peirce’s model of semiosis to a theory of culture structures (see also Wiley, 2006), their emergence, and their recursive institutionalization. These culture structures, in the line that links Peirce’s account of semiosis to Eco’s analysis of the encyclopedia, display however different properties that make them incompatible with the common way in which they are understood by cultural sociology.

First, their description is not dependent on a notion of semiotic ‘code’, as a set of fundamental oppositions and differences. Code theory usually stresses the boundedness of structures, their binary character, their paradigmatic form and (in some cases) their unconscious and universal nature (Heiskala, 2003: 191–200). Often as a consequence of the intersection of Durkheimian sociology and structural semiotics, these binary structures have been modeled upon the dichotomy between the sacred and the profane. But, unlike Durkheim, they give little room for both creativity and an account on the emergence of cultural structures. As a further point, they tend to favor a model of communication in which codes can automatically realize their communications, and in which communications are in turn easily decoded by interpreters. Norton (2014) criticizes this vision, and proposes to substitute the notion of ‘code’ with the notion of ‘cultural system’, a ‘system of relations in which meaning can be communicated and understood’, but which ‘does not, however, imply an efficient system’ (Norton, 2014: 171). The difference, in this case, is one that deals almost exclusively with the contingency of interpretation (unlike Norton, I tend not to introduce a break between interpretation and performance, and this is probably the most evident clue of the fact that he is still a neostructuralist at heart, while I am not). Efficiency, however, may be due to an unequal social distribution of sub-codes and connotations, thus being in great part resolved within the boundaries of code theory. If we speak of contingency, on the other hand, we make a decisive step toward a more accurate consideration of how ‘actors as causal intermediaries’ can enter the program of a ‘robust cultural sociology’ (Alexander, 2003: 26).

Second, if we assume that actors work concretely with and within culture in a way that is not dependent on codes as systems of relations among signifying units and as systems of rules, then we should address the problem of what is the alternative process. The concept of the encyclopedia is compatible with a vision of agents as involved in a continuous activity of selection and significance (Kockelman, 2013: 17–19) that involves different kinds of semiosic relations: of an object with a sign, and of a sign with its interpretant, as mediated by a selecting agent following the instigation of a significant object. This model (when fleshed out as a very precise and detailed meaning mechanism) would provide a better understanding of the organization and the realization of cultural structures. However, if we subscribe to this vision (and bring into it the kind of cognitive semiotics that Eco proposed when dealing with cognitive types and nuclear contents), we are once again leaving the canons of structural cultural sociology, because cultural autonomy is described as operating in substantially and radically different ways. To put it bluntly, the neo-structural version of cultural autonomy requires basic assumptions about the virtuality of structure. The kind of semiotic sociology that is drawn upon by Eco, on the contrary, rejects this duality of structure while placing duality at the level of the object (as immediate and dynamical object), all the while incorporating it into the same semiosic process. In terms of analytic reconstruction, this vision is more efficient because it does not duplicate structural components at any level of generality (at the level of binary oppositions, scripts, and performances), and places the very interpretability of objects through signs and interpretants at the center of how culture concretely works.

In truly triadic fashion, the first point would deal with the way signs are organized systemically (thus with an issue of ontology), while the second with the nature of objects and their relations to signs (thus highlighting the relevance of semantics). The third and final point, quite logically, regards the pragmatic openness of the contribution of Eco’s semiotics to social theory, and thus with the role of interpretants (in the broader encyclopedia and in the definition of content/meaning). The centrality of interpretants addresses issues that should be central both for the supporters of a strong symbolic vision of the cultural system, and for their critics who follow the non-representationalist path. The latter, while correctly arguing that some of culture is internalized without involving stable symbolic content, conflate symbolic activity and semiosis, neglecting the fact that some interpretants (while part of the semiosic process) instantiate procedures of affective orientation to an object and practical action upon it in the context-driven, and embodied manner they claim is the central dimension of culture-in-action. Mainstream cultural sociologists, on the other hand, would avoid both idealism and schematic transcendentalism in their account of culture, precisely because they would need to work with a notion of interpretation that formalizes the process of succession of interpretants in a way that accounts both for the creation and emergence of culture structures, and for their anchoring in real concrete semiotic processes, all characterized by the process of semiosis and involving not only symbols, but also habits, dispositions, and actions: as Eco claims,

the unlimited semiosis comes to an end, the exchange of signs has produced changes in experience, the missing link between semiosis and physical reality has finally been identified. The Theory of Interpretants is not idealistic. (Eco, 1979: 45)

I have made a case for the necessity to take this path and argued in favor of the usefulness of Umberto Eco’s semiotics (which I have tried to present extensively) as one of the possible references. There lies probably his relevance, not as an author who worked towards the development of an explicit sociological theory, but more in terms of a body of work that provides tools that can be integrated by social theorists if they want to pursue a different road toward the assessment of what makes cultural autonomy relevant.