Fundamentals of Semiotics

You already know how to follow an argument. You can spot a fallacy, reconstruct a proof, distinguish a valid inference from a plausible-sounding non sequitur. What you may not have is a systematic way of thinking about something more fundamental than argument—namely, meaning itself. How does a red octagon on a metal pole make you press the brake? How does a sequence of air-pressure variations leaving someone’s mouth become an idea in your head? How does a particular arrangement of fabric on your body tell strangers something about your social class before you open your mouth? These are not questions about logic, or psychology, or sociology in any straightforward sense. They are questions about signs—about how things come to mean—and the discipline that addresses them is semiotics.

Semiotics is, at its most compressed, the study of sign processes: how meaning is produced, structured, communicated, and transformed. But that compression hides what makes it genuinely useful. Semiotics gives you a vocabulary and a set of analytical mechanisms for taking apart something that ordinarily operates invisibly. You do not normally notice the semiotic machinery of a film poster, a medical symptom, a national flag, or a sentence in the way that you notice a headache. Meaning arrives pre-assembled, as if things simply are what they seem to mean. The central move of semiotics is to interrupt that transparency—to show that meaning is always constructed, always dependent on systems and conventions that could have been otherwise, and therefore always open to analysis.

The field has two founding figures who worked in complete independence of each other, on different continents, from different intellectual traditions, arriving at overlapping but importantly different frameworks. Ferdinand de Saussure (1857–1913), a Swiss linguist working in Geneva, approached signs from the standpoint of language and laid the groundwork for what became European structuralism. Charles Sanders Peirce (1839–1914; the surname is pronounced “purse”), an American philosopher, logician, and scientist, approached signs from the standpoint of logic and pragmatism. Saussure called his proposed science sémiologie; Peirce called his semiotic or semeiotic. The modern field uses “semiotics” as an umbrella term, but the two traditions differ in architecture, emphasis, and implication. Understanding both—and understanding where they differ—is essential equipment for thinking about meaning in any domain.

This primer takes Saussure as the primary path into the territory, for a practical reason: his framework is more tightly structured, easier to grasp as a system, and historically more influential in the humanities and social sciences. We will build up the Saussurean apparatus concept by concept, each one solving a problem raised by the previous one. Then we will introduce Peirce as an important alternative that expands the picture in directions Saussure’s model cannot reach. By the end of this first section, you should be able to do basic semiotic analysis of ordinary cultural objects using either framework—or both at once.

The problem that launched a discipline

Consider a thought experiment. You encounter the English word dog. You know what it means. But what, precisely, connects those three phonemes—the voiced alveolar stop, the open-mid back rounded vowel, the voiced velar stop—to the four-legged domesticated carnivore you are now picturing? The sounds themselves do not bark or wag. They bear no physical resemblance to canine anatomy. The connection is not causal in the way that smoke connects to fire, nor logical in the way that premises connect to conclusions. And yet the connection is robust, immediate, and shared across hundreds of millions of English speakers. Something is doing real work here, and that something is neither physics nor logic in any standard sense.

You might say: “The connection is just convention. English speakers agreed that dog means dog.” And you would be roughly right—but “convention” is not an explanation; it is a label for the thing that needs explaining. How does convention produce meaning? How do conventions get established, sustained, and modified? What is the internal structure of a conventional meaning-relation? If meaning is conventional rather than natural, what stops it from collapsing into chaos? And what about cases that seem partly conventional and partly not—a photograph that both resembles its subject and was physically caused by light bouncing off that subject, or a fever that both indicates and is partly constituted by disease?

Before Saussure, the dominant tradition in linguistics was historical and comparative. The Neogrammarians—the ruling school of nineteenth-century language study—traced how words evolved over centuries, how Latin became French and Italian and Spanish, how sound changes obeyed regular laws. This was rigorous and productive work, but it had a blind spot. It studied how languages change without first establishing what a language is. It catalogued individual elements—this word derives from that root, this sound shifted according to that law—without theorizing the system that gives those elements their function. Saussure’s pivotal insight was that you cannot understand a language by studying its parts in isolation any more than you can understand chess by studying the history of individual pieces. You must first understand the system of relations that makes the parts meaningful.

This is the foundational problem semiotics addresses: meaning is not a property of individual signs taken in isolation; it is a product of systems of relations. Everything that follows—in both Saussure and Peirce—is an attempt to specify what those systems look like, how they operate, and what they make possible.

The sign is not a label stuck on a thing

Saussure’s Course in General Linguistics was published in 1916, three years after his death. He never wrote the book. It was compiled by two colleagues, Charles Bally and Albert Sechehaye, from notes taken by students who attended his lectures at the University of Geneva between 1907 and 1911. This matters because the text is a reconstruction—probably more systematic and more tidy than Saussure’s actual thinking—and scholars have debated for a century how faithfully it represents his views. A manuscript discovered in 1996, later published as Writings in General Linguistics, provides some corrections and supplements. Nevertheless, the Course remains the founding document, and its conceptual architecture has proven extraordinarily durable.

Saussure begins by attacking what he calls the “nomenclaturist” view of language—the intuitive picture that language is essentially a list of names, each attached to a corresponding thing in the world. On this naive view, learning French means learning that the thing you call dog the French call chien, and learning Mandarin means learning that they call it gǒu. Language is a labeling system; the world comes pre-divided into categories, and each language merely pastes different labels on the same categories.

Saussure saw that this picture fails on multiple fronts. It assumes that “ready-made ideas exist before words.” It cannot explain why different languages carve up experience in different ways (a point we will return to). And it treats meaning as a simple one-to-one mapping between words and things, which cannot account for abstract terms, relational terms, negation, or fiction—you can talk about unicorns, absent friends, and the square root of negative one despite the absence of corresponding objects to stick labels on.

In place of the nomenclaturist picture, Saussure proposes the sign (signe), defined as the union of two components:

The signifier (signifiant): what Saussure calls the “sound-image” (image acoustique). This is not the physical sound wave that travels through the air. It is the psychological impression of the sound—the mental pattern you can run through your head when you silently recite a poem or rehearse what you plan to say. Saussure demonstrates the point simply: “Without moving our lips or tongue, we can talk to ourselves or recite mentally a selection of verse.” What we are manipulating in inner speech is the signifier—a psychological entity, not an acoustic one.

The signified (signifié): the concept associated with the signifier. Crucially—and this is among the most frequently misunderstood points in all of semiotics—the signified is not the referent. It is not the actual flesh-and-blood dog, the physical tree, or the historical Napoleon. It is a mental concept: the idea of dog-ness, tree-ness, or Napoleon. The word unicorn has a perfectly functional signified (you can think the concept “unicorn” without difficulty) despite having no referent in the physical world. Saussure deliberately brackets the referent—excludes it from his model—because he wants to analyze the internal structure of the sign system without getting tangled in questions about the relationship between language and external reality. The philosopher Susanne Langer put the point with useful clarity: “Symbols are not proxy for their objects but are vehicles for the conception of objects. It is the conceptions, not the things, that symbols directly mean.”

This is worth pausing on because the nomenclaturist error is deeply intuitive and persistent. When people first encounter the signifier/signified distinction, they almost invariably understand the signifier as “the word” and the signified as “the thing the word refers to.” But both components are psychological. The sign is an association between a mental sound-pattern and a mental concept, residing entirely within the minds of language users. It is, as Saussure puts it, “like two sides of a piece of paper”—you cannot cut one side without cutting the other, and neither side exists independently.

The practical consequence is immediate. When you see the word OPEN on a shop door, the sign consists of the sound-image /ˈoʊpən/ associated with the concept available for business. When you see the same word OPEN on an elevator button, the signifier is identical but the signified has shifted: now it means cause the doors to part. Same signifier, different sign—because the sign is the union of both components, not either one alone. And an outline of a box with an open flap on a shipping carton? Different signifier entirely, yet a related signified. Signs are not fixed connections between sounds and things; they are structured pairings of mental entities, and those pairings can shift, overlap, and multiply.

Why the sign is arbitrary—and why that matters enormously

Saussure’s next major claim is the principle of arbitrariness: “The bond between the signifier and the signified is arbitrary.” The idea of “sister” is not linked by any inner relationship to the French sound-sequence sœur; the same concept is equally well served by sister, Schwester, hermana, or shimai. The very existence of different languages proves the point. There is no natural, necessary, or logical connection between any particular sound-pattern and any particular concept.

But Saussure is careful to specify what “arbitrary” does not mean. It does not mean random, meaningless, or freely chosen by individual speakers. “The word arbitrary should not imply that the choice of the signifier is left entirely to the speaker,” he writes. “I mean that it is unmotivated, i.e., arbitrary in that it actually has no natural connection with the signified.” Once a sign is established in a linguistic community, no individual can change it by fiat. Language is inherited; you learn the conventions of your speech community, and those conventions constrain you absolutely in everyday communication. The sign is arbitrary in its origin (there is no reason why these sounds should carry this meaning rather than those) but obligatory in its operation (once established, speakers are bound by the convention). The anthropologist Claude Lévi-Strauss compressed this into a useful formula: the sign is “arbitrary a priori but ceases to be arbitrary a posteriori.”

Apparent counterexamples arise immediately. What about onomatopoeia—words like splash, cuckoo, buzz, or tick-tock that seem to imitate their referents? Saussure addresses this directly. First, he notes that genuinely onomatopoeic words are far fewer than people suppose. Second, even apparent onomatopoeia varies across languages: a rooster says cock-a-doodle-doo in English, cocorico in French, kikeriki in German, and kokekokkō in Japanese. If the sound-meaning connection were natural, these words should be identical. Third, once onomatopoeic words enter the language, they undergo the same historical sound changes as all other words, gradually losing their mimetic quality. The French word pigeon derives from Vulgar Latin pipio, which was originally onomatopoeic—but no modern French speaker hears the cooing of a bird in the word pigeon. Interjections receive the same treatment: French aïe! and English ouch! express the same concept (pain) with different signifiers—proof that even these seemingly natural expressions are conventionally shaped.

Why does arbitrariness matter so much? Because it entails that meaning is not grounded in nature but in convention—and therefore in social, cultural, and historical processes that could have been otherwise and can change. If the relationship between sound and concept were natural, meaning would be fixed, universal, and unchallengeable. Arbitrariness means that meaning is fundamentally a social product: it requires a community of users who sustain the convention. This makes meaning available for cultural analysis in a way that natural connections would not be. The price of a word, like the price of a currency, is set by the system in which it circulates, not by any intrinsic property of the token itself.

Arbitrariness also explains language’s extraordinary expressive power. If signs had to resemble what they represent, language would be limited to things that can be acoustically imitated—concrete, sensory, present. Abstract concepts like justice, possibility, negation, and the square root of negative one would be unsignifiable. Because the link between sound and concept is unmotivated, any sound-pattern can in principle be assigned to any concept, giving language an essentially unlimited representational range.

The most philosophically important consequence, however, is one that Saussure draws out in his discussion of value—a concept we turn to now, and the hardest single idea in Saussurean semiotics.

Meaning lives in the gaps between signs

If signs are arbitrary—if there is no natural bond between signifier and signified—then what does determine a sign’s meaning? Saussure’s answer is his most radical claim: “In language there are only differences. Even more important: a difference generally implies positive terms between which the difference is set up; but in language there are only differences without positive terms.” This is the thesis that meaning is purely differential and relational—that a sign means what it means not because of any inherent content but because of how it differs from other signs in the same system.

To unpack this, Saussure distinguishes between signification and value (valeur). Signification is the internal relationship within a single sign—the pairing of a particular signifier with a particular signified. It is what you find in a dictionary entry. Value is something else: it is the sign’s position within the entire system of signs, determined by its relations of contrast and difference with every other sign.

Saussure’s own example makes this vivid. The French word mouton and the English word sheep have the same signification—both can refer to the woolly animal in the field. But they do not have the same value. English has two words where French has one: sheep for the living animal and mutton for the meat on the plate. The existence of mutton in the English system constrains the value of sheep, restricting it to the animal. French mouton, facing no such competitor, covers both animal and meat. The value of sheep is not determined by pointing at an animal and saying “that”; it is determined by the system of distinctions within English—specifically, by the presence of mutton beside it.

An analogy Saussure favors: think of a five-franc coin. To understand what it is “worth,” you need two things. First, you need to know what it can be exchanged for—a certain quantity of bread, say. This is analogous to signification (the sign’s internal pairing of signifier and signified). Second, you need to know how it compares to other coins in the same monetary system—to one-franc pieces, ten-franc pieces, and so on. This is analogous to value (the sign’s position in the system). And here is the crucial point: “It is not the metal in a coin that fixes its value.” A coin’s purchasing power is not a physical property of the disc; it is a function of the monetary system. Likewise, a sign’s meaning is not an intrinsic property of the sounds; it is a function of the sign system.

The consequences are profound. Consider the color spectrum. Physically, the electromagnetic spectrum of visible light is a continuous gradient with no inherent boundaries. Yet different languages divide this continuum differently. English treats blue and green as fundamentally different colors. Russian makes an obligatory distinction that English does not: goluboy (light blue) and siniy (dark blue) are not “shades of blue” in Russian but distinct basic color categories—as different as green and blue are in English. A landmark 2007 study published in the Proceedings of the National Academy of Sciences found that Russian speakers were measurably faster at discriminating colors that fell across the goluboy/siniy boundary compared to colors of equal physical difference within a single category, while English speakers showed no such advantage. When verbal interference was introduced (subjects had to simultaneously rehearse a string of digits), the Russian advantage disappeared—demonstrating that the linguistic distinction was actively shaping perception in real time.

Welsh provides another case: the word glas covers what English would call blue, green, and sometimes grey, depending on context. Japanese ao historically covered both blue and green; to this day, traffic lights in Japan are called ao even though the “go” light is unambiguously green. Hungarian treats bright red (piros) and deep red (vörös) as different basic colors, not mere shades. These are not failures of color vision. They are demonstrations of Saussure’s point: language does not label a pre-existing reality; it helps carve up reality into the categories through which we experience it. The value of English blue is determined by the fact that green exists beside it, and purple exists on the other side. Remove green from the system, and the value of blue would expand to fill the gap, as it does in languages that lack the distinction.

Saussure’s chess analogy crystallizes the general principle. A chess bishop is a bishop not because of any inherent property of the carved piece of wood. If the bishop is lost, you can replace it with a button, a bottle cap, or a folded napkin—anything will do, as long as it is distinguishable from the other pieces. What makes it a bishop is its differential function within the system: it moves diagonally, it is not a rook, it is not a knight, its value is defined entirely by the rules and by its contrast with the other pieces. Similarly, Saussure writes, “each linguistic term derives its value from its opposition to all the other terms.” Language is, in his compressed formulation, “a system of pure values.”

This is the moment where Saussure’s framework stops being a set of definitions and becomes a genuine intellectual tool. The system-of-differences thesis means that you can never fully understand a sign by examining it in isolation. You must always ask: what does this sign contrast with? What other signs occupy neighboring positions in the system? What would change if one of those neighbors were added, removed, or shifted? These questions are the engine of semiotic analysis, and they apply far beyond language. The value of a business suit is determined not by the thread count of its fabric but by the existence of jeans, hoodies, and tuxedos in the same vestimentary system. The value of a horror film is defined partly by its contrast with thrillers, comedies, and dramas. Wherever there is a system of meaningful alternatives, Saussurean value is operative.

The system behind every utterance

Saussure draws a sharp distinction between langue and parole that is essential to his method and, like most Saussurean distinctions, frequently misunderstood. Langue is the abstract system of a language—the complete inventory of signs, the rules for combining them, the network of differential relations that gives each sign its value. It is shared across a speech community, stored (as Saussure puts it) “in the brain” of each speaker as a collective deposit, something like a shared dictionary and grammar that every competent speaker has internalized. Parole is actual speech—the concrete, individual, variable utterances that real people produce in real situations, complete with hesitations, errors, accents, innovations, and contextual nuances.

The distinction matters because Saussure needs to carve out a proper object for scientific study. Actual speech is a bewildering mess. People mumble, stutter, use words in novel ways, talk over each other, coin slang, and violate grammatical rules. If you tried to do science on the totality of all actual English utterances, you would drown in heterogeneity. Langue provides a homogeneous, systematic object—the underlying structure that makes all those variable utterances possible and intelligible. Parole is the data; langue is the system that generates and organizes the data.

An analogy from music helps. Every actual performance of a Beethoven symphony is parole—different tempos, different volumes, different concert halls, different moments of interpretation by the conductor. The score is closer to langue: the abstract system of relations (notes, rhythms, dynamics, instrumentation) that every performance realizes in a different way. You study the score to understand the symphony’s structure; you listen to performances to understand its life. Saussure argues that linguists should focus on langue, at least as a methodological starting point, because it is the systematic component—the part amenable to structural analysis.

This parallels a distinction familiar from analytic philosophy. Noam Chomsky later proposed the distinction between competence (a speaker’s internalized knowledge of their language) and performance (the actual use of language in concrete situations), and explicitly acknowledged the parallel with Saussure. But the two are not identical. Saussure’s langue is fundamentally social—it is a shared, communal system, a “social product” deposited in each brain through participation in a speech community. Chomsky’s competence is fundamentally individual and psychological—the internalized grammar of an ideal speaker-listener, grounded in an innate language faculty. Saussure treats sentence construction as belonging to parole (the domain of individual combination); Chomsky places the generative capacity to produce infinite sentences squarely within competence. These are genuinely different theoretical commitments, not just terminological preferences, and they lead to different research programs. But the shared structural insight—that you must distinguish the system from its use in order to do productive analysis—is the critical takeaway.

One feature of langue that beginners sometimes miss: it is not a prescriptive grammar. Langue is not the set of rules your English teacher told you to follow. It is the complete set of conventions that a speech community actually operates with, including informal registers, slang, and constructions that prescriptivists would condemn. If most English speakers understand and use “Who did you give it to?” (ending a sentence with a preposition), then that construction is part of English langue, regardless of what any style guide says. Langue is descriptive, not normative. It is the system as it actually functions, not as anyone thinks it ought to function.

Two axes of meaning: combination and selection

The final major component of the Saussurean framework is the distinction between syntagmatic and paradigmatic relations—two axes along which every sign simultaneously operates.

Syntagmatic relations are relations of combination and sequence. They govern how signs chain together in linear order. In a sentence, each word stands in a syntagmatic relationship with the words before and after it: “The cat sat on the mat” is a syntagm. The relations are in praesentia—the connected signs are all present together in the actual utterance. Word order matters; “the mat sat on the cat” is a different syntagm with a different (and strange) meaning. Grammar is largely a system of syntagmatic constraints: rules about what can combine with what, and in what order.

Paradigmatic relations are relations of selection and substitution. They connect each sign to the set of alternatives that could have filled the same position. In “The cat sat on the mat,” the word cat belongs to a paradigmatic set that includes dog, child, spider, philosopher—anything that could grammatically and meaningfully replace it in that slot. The word sat belongs to a paradigm including lay, slept, died, danced. These relations are in absentia—the alternatives are not present in the utterance, but they exist as a virtual set in the background of the language system. Saussure called them “associative relations”; the term “paradigmatic” was introduced later by the linguist Roman Jakobson and became standard.

The value of any sign is determined by both axes simultaneously. On the syntagmatic axis: what does the sign combine with? On the paradigmatic axis: what does the sign contrast with? Change either, and you change the meaning. Replace cat with assassin and the sentence means something very different—the paradigmatic substitution has altered the whole syntagm’s significance. Keep cat but change the syntagm to “The cat devoured the mat” and the new syntagmatic context shifts what we understand about the cat.

These axes apply far beyond language. Roland Barthes, who did more than anyone to extend Saussure’s framework to culture at large, applied them brilliantly to food and clothing. Consider a restaurant meal. The sequence of courses—starter, main, dessert—is a syntagm: a structured combination in a conventional order. At each position in the sequence, the diner faces a paradigmatic choice: for the starter, you select one from a set of alternatives (soup, salad, bruschetta). The meaning of your meal—what it communicates about your taste, your budget, your cultural positioning—arises from the intersection of both axes: which items you selected (paradigm) and how they combine in sequence (syntagm).

Clothing works the same way. Getting dressed involves a series of paradigmatic choices within functionally defined categories: from the set of “tops,” you choose a shirt, blouse, or sweater; from “bottoms,” trousers, skirt, or shorts; from “footwear,” boots, sneakers, or heels. The outfit as a whole is a syntagm—a combination that must cohere according to conventional rules. A tailored blazer with swim trunks and hiking boots creates a syntagmatic clash, the vestimentary equivalent of a grammatically broken sentence. It either fails to communicate coherently or communicates deliberate rule-breaking—which is itself a meaningful semiotic act, legible precisely because everyone recognizes the rules being violated.

One more illustration of how paradigmatic relations work, because this concept tends to be underappreciated. When a newspaper editor chooses a photograph to accompany a political story, the choice is paradigmatic: from a set of available images (the politician smiling, the politician frowning, the politician looking confused, the politician gesturing forcefully), one is selected. The selected image is visible in the paper; the rejected alternatives are invisible. But the meaning of the chosen image is partly constituted by the alternatives that were not chosen. Publishing the frowning photo means something different if the smiling photo was also available—the omission is as significant as the inclusion. Every paradigmatic choice is simultaneously an exclusion, and semiotic analysis frequently requires asking not just “what signs are present?” but “what signs could have been here but aren’t?”

A different architecture of the sign

We have now assembled the core Saussurean toolkit: the sign as a dyadic union of signifier and signified, the arbitrariness of their bond, the system of differences that gives each sign its value, the distinction between the system (langue) and its use (parole), and the two axes of combination (syntagmatic) and selection (paradigmatic). This is a powerful analytical framework, and it has driven productive work across the humanities for over a century. But it has limitations—limitations that become visible when we encounter the alternative framework developed independently by Charles Sanders Peirce.

Peirce was not a linguist. He was a philosopher, logician, and practicing scientist (he worked for thirty-two years at the U.S. Coast and Geodetic Survey). He is the founder of pragmatism, the distinctively American philosophical tradition that ties the meaning of a concept to its conceivable practical effects. For Peirce, semiotics was not a subdiscipline of linguistics but a branch of logic—indeed, he held that “logic, in its general sense, is only another name for semiotic.” Where Saussure derived his sign theory from the study of language and proposed extending it to other sign systems, Peirce began from the most general possible question: what does it mean for anything to stand for anything else to anyone?

The most immediately visible difference is structural. Where Saussure’s sign is dyadic (two components: signifier and signified), Peirce’s sign is triadic (three components). Peirce argued that semiosis—the process of meaning-making—is fundamentally a three-term relation that cannot be reduced to any combination of two-term relations. His canonical formulation: “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.”

The three components:

The representamen (or sign-vehicle) is the perceptible form that does the representing—the thing you encounter that functions as a sign. In practice, this roughly parallels Saussure’s signifier, though Peirce conceives it more broadly: it need not be linguistic or even man-made. Smoke, a footprint, or a flushed face can serve as a representamen.

The object is what the sign represents or refers to. Unlike Saussure, who deliberately excluded the referent from his model, Peirce explicitly includes the object as a structural component of the sign. He later refined this into the immediate object (the object as the sign represents it—always partial and perspectival) and the dynamic object (the actual object in its full reality, which no single sign can fully capture). This distinction acknowledges that signs always present their objects under a certain aspect or from a certain angle, never exhaustively.

The interpretant is the most innovative component, and the one with no Saussurean equivalent. The interpretant is the effect the sign produces—the understanding, response, or “translation” it generates. It is the meaning the sign gives rise to in the mind of whoever encounters it. A critical clarification: the interpretant is not the interpreter (the person). It is the interpretation-effect—the thought, reaction, or further sign that the encounter with the sign produces. Peirce later distinguished the immediate interpretant (the interpretive potential inherent in the sign itself), the dynamic interpretant (the actual effect produced on a particular occasion), and the final interpretant (the effect that would be produced under ideal conditions of interpretation and inquiry).

Why does the triadic structure matter? Because it builds interpretation into the very definition of the sign. In Saussure’s model, the sign is a structural entity—a pairing of signifier and signified that exists within the system of langue—and interpretation is, in a sense, taken for granted. The system produces meaning; speakers access the system; meaning arrives. In Peirce’s model, there is no sign without an interpretant: something only functions as a sign insofar as it produces an interpretation. Smoke in an empty forest, observed by no one and generating no interpretant, is not yet functioning as a sign. A fever that no one notices is not yet semiotically operative. Peirce’s framework foregrounds the process of interpretation as constitutive of meaning, not secondary to it.

Icons, indices, and symbols: three ways a sign can mean

Peirce was a compulsive taxonomist (he eventually developed a scheme of sixty-six classes of signs, which even he admitted he found uncertain), but his most famous and widely used classification is the trichotomy of icon, index, and symbol—distinguished by the nature of the relationship between the sign and its object.

An icon signifies by virtue of resemblance. The sign shares some quality or structural feature with its object. A portrait is an icon of the person it depicts: it looks like them. A map is an icon of the territory: its spatial relations correspond to spatial relations on the ground. An architect’s blueprint is iconic: the proportions on paper mirror the proportions of the building. Onomatopoeia is iconic in the auditory domain: buzz sounds (somewhat) like the sound it represents. Peirce subdivided icons into three types: images (which share simple qualities with their objects—a color swatch matching a paint color), diagrams (which share relational structures—an organizational chart representing a corporate hierarchy), and metaphors (which represent through a parallelism between two domains—“life is a journey” maps the structure of travel onto the structure of a lived experience).

An index signifies by virtue of a causal, physical, or existential connection to its object. Smoke is an index of fire: it is caused by fire, and encountering smoke directs your attention to fire as its source. A bullet hole is an index of a gunshot. A weathervane is an index of wind direction—the wind physically moves it. A fever is an index of infection: the illness causes the elevated temperature. A knock on the door is indexical—it draws attention to someone’s presence. Peirce considered indexicality essential: “Anything which focusses the attention is an index.” Demonstrative pronouns (this, that, here), proper names, and pointing gestures are all indexical—they single out particular objects by directing attention toward them rather than describing them. Peirce came to see indices as indispensable to logic itself, because without them, signs would remain entirely general and abstract, unable to refer to particular existing things.

A symbol signifies by virtue of convention, habit, or social agreement. Most words in any language are symbols: there is no resemblance between the word justice and the concept of justice, no causal connection between the sound tree and arboreal objects. The connection is established and maintained by convention within a community of users. National flags are symbols: the specific arrangement of colors and shapes represents a nation by agreement, not by resemblance or causation. Wedding rings, traffic signals, mathematical notation, and corporate logos are all primarily symbolic.

Now, the crucial point that beginners almost universally miss: these three categories are not mutually exclusive. By 1903, Peirce was explicit that “it would be hard, if not impossible, to find any pure instances of icons and indices.” Almost every real sign participates in multiple categories simultaneously. A photograph is the paradigm case: it is iconic (it visually resembles its subject), indexical (it was physically produced by light reflecting off the subject and striking the film or sensor—a causal chain), and partly symbolic (conventions of framing, composition, and genre shape how we read it). This triple character is precisely what gives photographs their peculiar epistemic authority—they feel more “truthful” than paintings because they combine resemblance with a causal trace, an indexical guarantee that “this was really there.”

A medical symptom like a rash is indexical (caused by the underlying condition) but may also be iconic (a measles rash has a characteristic appearance that resembles—and thereby identifies—measles specifically) and symbolic (within clinical practice, the rash is read through a conventional diagnostic code). A no-smoking sign showing a cigarette silhouette crossed by a red slash is iconic (the silhouette resembles a cigarette), symbolic (the red circle-and-slash means “prohibited” by convention), and indexical (it is placed at locations where smoking is forbidden, connecting it existentially to a specific rule in a specific place). Rather than classifying signs as exclusively one type, skilled semiotic analysis asks: in what respects is this sign iconic, indexical, and symbolic? What work is each dimension doing?

This is where Peirce’s framework reveals something that Saussure’s cannot easily capture. Saussure treats all linguistic signs as essentially arbitrary (symbolic in Peirce’s terms) and focuses on the internal system of differences within language. This is powerful for analyzing linguistic structure but leaves it hard to analyze signs that operate through resemblance or causal connection. Peirce’s icon/index/symbol trichotomy gives you a way to distinguish and analyze the different mechanisms by which any sign—linguistic or otherwise—connects to what it represents. A traffic sign, a medical image, a social media emoji, a forensic photograph, and an algebraic equation all work differently, and Peirce’s framework lets you specify how they differ.

Meaning never sits still

The deepest difference between Saussure and Peirce is not structural (dyad versus triad) but dynamic. Saussure’s model is fundamentally synchronic and structural: it describes the system of signs as it exists at a given moment, a static architecture of differences. Peirce’s model is fundamentally processual: meaning is not a fixed state but an ongoing process he calls semiosis.

Semiosis works like this. You encounter a sign (representamen) that represents an object. This encounter produces an interpretant—an understanding, a response, a translation. But that interpretant is itself a sign: it is a new representation that stands in relation to the same object. This new sign in turn produces its own interpretant, which is again a sign, which produces yet another interpretant—and so on. Each act of interpretation generates a further sign, which becomes the starting point for further interpretation. Peirce captured this with characteristic vividness: “The interpretant is nothing but another representation to which the torch of truth is handed along; and as representation, it has its interpretant again. Lo, another infinite series.”

The Italian semiotician Umberto Eco later coined the term unlimited semiosis to describe this Peircean principle: meaning is never finally and completely determined, because every interpretation can in principle be further interpreted, refined, challenged, or extended. When you look up a word in the dictionary, the definition is given in more words, which could themselves be looked up, each definition generating further definitions in an endless chain. When you explain to someone what a film “means,” your explanation is a new sign of the same object (the film’s meaning), which your interlocutor interprets in their own way, producing yet another sign—and the conversation can continue indefinitely.

This might seem to imply that meaning is infinitely unstable, that “anything goes,” that interpretation is arbitrary. But Peirce explicitly blocks this conclusion. The object constrains the sign: not every interpretation of smoke is a valid one (it is an index of fire, not of ice cream). Peirce’s pragmatism further anchors meaning: the meaning of a concept consists in its conceivable practical effects, and these provide a regulative standard. Unlimited semiosis is not unstructured free association. It is a directed process—constrained by the object, by the sign’s formal properties, and by the habits and conventions of the interpretive community—that is nonetheless always open to further development. Meaning is not fixed and eternal, nor is it chaotic and arbitrary. It is stable enough to enable communication and action, but open enough to permit growth, revision, and discovery.

This dynamic quality is what makes Peirce’s framework particularly valuable for analyzing processes of cultural change, reinterpretation, and contestation. When a political symbol is reclaimed by a group that was originally stigmatized by it—when a slur is appropriated and resignified, or a corporate logo becomes the target of protest—these are semiotic processes in Peirce’s sense: the same representamen, referring to the same object, generating new and different interpretants across different communities and historical moments. Saussure’s synchronic system of differences can describe the state of the sign system at any given moment but has difficulty theorizing the dynamics of change. Peirce’s processual model is built for it.

Seeing both frameworks in action

Let us put both toolkits to work on a single concrete case: a print advertisement. Imagine a luxury watch ad in a glossy magazine. The photograph shows a man in an expensive suit, standing on a yacht, wearing the watch. Below: the brand name, a tagline, and a price.

A Saussurean analysis would begin by identifying the signs and their relations. The brand name (signifier: the graphic form of the name; signified: the concept of the brand and its associations) is a sign whose value is determined by its position in the system of luxury watch brands—it means what it means partly by not being Rolex, Omega, or Casio. The image of the yacht is a signifier whose signified is not “this specific yacht” but the concept of wealth, leisure, maritime freedom. The suit signifies professionalism and success, its value determined by contrast with what it is not (casual wear, workwear, athletic wear). The syntagmatic structure matters: the combination of man + suit + yacht + watch + brand name forms a coherent “sentence” whose elements reinforce each other. Substitution on the paradigmatic axis reveals the ad’s choices and their significance: why a yacht and not a sports car? Why a suit and not a tuxedo? Each paradigmatic alternative that was rejected is part of the ad’s meaning.

A Peircean analysis would ask different questions. The photograph is simultaneously iconic (it looks like a man on a yacht), indexical (as a photograph, it bears a causal trace—light from the scene struck the sensor), and symbolic (the conventions of luxury advertising shape how we read it). The watch on the man’s wrist is an index of his status (he possesses it, he wears it, it connects him to the brand). The brand logo is primarily a symbol (conventional association), but if the logo’s design incorporates a crown or shield, those elements are iconic (resembling emblems of royalty and heraldry). The tagline, if it exists, is symbolic (conventional language). The entire ad functions as a complex sign whose object is not the physical watch but the brand-as-value-system, and whose interpretant is the viewer’s response: desire, aspiration, recognition of status, or skepticism and critique. That interpretant—whatever it is—becomes a further sign (perhaps a conversation with a friend about whether luxury watches are worth it), generating further interpretants, in an ongoing process of semiosis.

Notice what each framework illuminates. Saussure gives you the system of differences—the paradigmatic and syntagmatic structure that makes the ad’s specific choices meaningful. It shows you the architecture. Peirce gives you the mechanisms of signification—how different parts of the ad connect to their objects through resemblance, causal trace, and convention—and the dynamic process by which the ad produces effects in viewers that then circulate socially. It shows you the machinery in motion. The frameworks are not competitors; they are complementary lenses that focus on different aspects of the same phenomenon.

The canonical applied example in semiotic literature is Roland Barthes’s 1964 analysis of a Panzani pasta advertisement—an image showing packets of branded pasta and a tin alongside fresh tomatoes, peppers, mushrooms, and onions, all spilling from a half-open string shopping bag against a red-and-green background. Barthes identifies multiple layers of signification operating simultaneously. The string bag connotes freshness and market shopping (as opposed to supermarket plastic). The tumble of ingredients connotes abundance and the promise of a home-cooked meal. The red, green, and yellow colors echo the Italian flag, connoting “Italianicity”—authenticity, tradition, Mediterranean vitality. Even the brand name Panzani participates: it sounds Italian (the double consonant, the vowel ending), connoting the same complex of associations. Barthes distinguishes the denotative level (these are recognizably tomatoes, pasta, peppers) from the connotative level (these signs carry cultural associations of freshness, authenticity, domesticity, Italianness) and argues that connotation is where ideology operates—where culturally specific values are made to seem natural and self-evident. This analysis, which we will explore more deeply in the next section, demonstrates how semiotics moves from a set of abstract principles to a concrete critical practice.

What comes next—and why it matters

The conceptual machinery assembled here—the sign, arbitrariness, value, system of differences, langue/parole, syntagmatic/paradigmatic axes, icon/index/symbol, semiosis—is foundational. It gives you the vocabulary and the analytical reflexes to begin thinking semiotically about any domain where meaning is at stake. But it is also, deliberately, a ground floor. Several essential topics remain.

Codes and conventions formalize how signs are organized into systems beyond language: visual codes, narrative codes, behavioral codes. Without a theory of codes, you cannot explain why certain signs are legible in some contexts and opaque in others, or how subcultures create meanings that are invisible to outsiders.

Denotation and connotation, briefly mentioned through Barthes’s pasta ad, need systematic treatment. The mechanism by which a first-order sign (denotation) becomes the signifier for a second-order meaning (connotation) is Barthes’s most important contribution, and it leads directly to the concept of myth—the process by which culturally specific, historically contingent meanings are made to seem natural, universal, and innocent. Myth is where semiotics becomes a tool of ideology critique, connecting the analysis of signs to questions of power.

Ideology and discourse extend the framework further, incorporating the work of Louis Althusser and Michel Foucault, to ask how sign systems do not merely reflect social reality but help construct and sustain it. The signs that seem most “natural”—gender categories, racial categories, the visual language of advertising, the narrative structures of journalism—are precisely the ones doing the heaviest ideological work.

Finally, specialized applications in visual semiotics, digital media, architecture, music, and branding demonstrate that the principles developed here for language extend, with modifications, to every domain of human meaning-making. The second part of this primer will build out these tools, moving from the foundations established here toward the analytical capacity to take apart complex cultural texts and understand how they do their work.

For now, the essential insight to carry forward is this: meaning is not a mysterious property that some things just “have.” It is a product of structured systems and dynamic processes that can be analyzed with precision. Signs are not transparent windows onto a pre-given reality; they are constructed, conventional, and differential. Understanding how they work does not diminish their power—it reveals the machinery that makes their power possible. That machinery is what semiotics teaches you to see.

Codes: the rules that make reading possible

The sign, whether you build it from Saussure’s two-part model or Peirce’s three-part model, never operates alone. Signs always appear inside larger systems of organization—rules that tell you which signifiers go with which signifieds, which combinations are permitted, which are forbidden, and what counts as a meaningful difference versus an irrelevant variation. These organizing systems are called codes, and understanding them is the bridge between knowing what a sign is and being able to analyze how signs actually work in culture.

A code, in semiotic terms, is a set of conventions that correlates signifiers with signifieds, enabling the production and interpretation of meaning. Think of it this way: a single red octagon, absent any code, is just a shape with a color. Within the code of traffic signs, it means STOP. Within the code of geometry, it means octagon. Within the code of a particular video game, it might mean health power-up. The sign does not carry its meaning around like a suitcase. The code assigns the meaning, and different codes assign different meanings to the same material form.

This insight has enormous analytical consequence. It means that to interpret any sign correctly, you must already possess the relevant code—and that possession of codes is unevenly distributed across populations. Codes are social. They are learned, not innate. They are maintained by communities of users. And they are enforced, sometimes formally (law, grammar), sometimes informally (ridicule, confusion, exclusion). Stuart Hall put it directly: “There is no intelligible discourse without the operation of a code.”

The relationship between codes and Saussure’s foundational concept of langue is close but not identical. Saussure’s langue was the abstract system of a single natural language—French, English, Mandarin—shared by a speech community. Roman Jakobson explicitly substituted the pair code/message for Saussure’s langue/parole, broadening the concept beyond verbal language to include any rule-governed sign system. The code concept lets you apply structural analysis to visual communication, body language, fashion, architecture, ritual behavior, and digital interfaces—any domain where conventions correlate forms with meanings.

Types of codes and their domains

Codes operate across every domain of human communication, and cataloguing them helps reveal just how thoroughly coded everyday life is. Linguistic codes include phonological rules (which sound combinations are legal in a language), syntactic rules (how words combine into sentences), and pragmatic conventions (how context shapes utterance meaning). Visual codes govern how images communicate: color symbolism, compositional conventions, iconographic traditions, the “grammar” of camera angles in photography and film. Behavioral codes regulate interaction: the unspoken rules of personal space, eye contact, turn-taking in conversation, the rituals of greeting and departure. Cultural codes are broader organizing frameworks—the shared assumptions, values, and knowledge structures that a community uses to make sense of experience. And textual codes are genre conventions that tell you how to read a particular kind of text: the codes of detective fiction, news reporting, academic writing, or Instagram posts each establish expectations about form, content, and appropriate interpretation.

What makes codes analytically powerful is their dual function: they simultaneously enable and constrain interpretation. Codes enable by providing the shared framework that makes communication possible at all—without the code of English, this sentence is noise. But codes also constrain by limiting the range of meanings a sign can plausibly carry. A red traffic light, within its code, cannot mean “accelerate.” Textual codes constrain more subtly: within the code of realist fiction, a character’s sudden ability to fly would violate genre expectations and produce confusion or comedy, not seamless narrative. Codes channel interpretation toward certain meanings and away from others, and this channeling is never ideologically innocent—a point we will return to repeatedly.

Codes also create boundaries between insiders and outsiders. Knowledge of a code marks membership in a community; ignorance of it marks exclusion. Medical jargon, legal terminology, the argot of subcultures, the visual codes of fashion tribes, the interaction protocols of online communities—all function as gatekeeping mechanisms. Dick Hebdige, following Hall, analyzed how punk subculture deliberately constructed an oppositional code system—safety pins, torn clothing, aggressive typography—that simultaneously communicated meaning within the group and signaled defiant otherness to those outside it. The subcultural code was legible to insiders and deliberately illegible, or offensive, to the mainstream. Every professional community does something similar, if less theatrically: the codes of academic writing, for instance, mark disciplinary belonging while excluding non-specialists.

Eco’s theory of codes: precision about machinery

Umberto Eco, in A Theory of Semiotics (1976), brought unusual precision to code theory by distinguishing between s-codes and codes proper. An s-code (the “s” stands for system) is a structural system—an organized set of elements defined purely by their internal relationships and oppositions. The set of traffic light colors (red, yellow, green) is an s-code on the expression plane. The set of traffic commands (stop, caution, go) is another s-code on the content plane. Each element gets its identity from its position relative to the others, exactly as Saussure argued for linguistic signs: red means what it means because it is not yellow and not green.

A code proper, in Eco’s terminology, is the correlational rule that maps elements from one s-code to elements of another. It is the convention that connects expression to content—the rule that says red light corresponds to stop. The code is not the system itself but the bridge between two independent systems. This distinction matters because s-codes can exist without being correlated to anything (the phonological system of a language exists as a structure before its sounds are assigned to meanings), whereas a code proper always involves correlation between expression and content planes.

Eco then introduced three operations that describe what happens when codes meet actual interpretation. Overcoding occurs when supplementary rules are added on top of existing codes, narrowing interpretation further. Stylistic conventions in literature are overcodes: they take the base linguistic code and add extra rules about register, genre expectations, and rhetorical form. When you recognize that a sentence is written in the style of hard-boiled detective fiction rather than Romantic poetry, you are applying overcodes that restrict meaning beyond what the base linguistic code specifies. Undercoding is the opposite operation—a tentative, provisional interpretation when established codes are unclear or absent. When you encounter an unfamiliar sign system—a foreign culture’s gestures, an avant-garde artwork, an ambiguous emoji—you hypothesize codes to make sense of what you see. Undercoding produces approximate, revisable interpretations rather than precise decodings. The third operation, extracoding, is the broader category encompassing both: any interpretive work that goes beyond what a given code strictly provides.

These concepts solve a real problem. Simple code models, inherited from information theory, treat coding as mechanical substitution—one element neatly replaced by another, like a cipher. Eco argued this model fails to capture what actually happens in interpretation. Even apparently mechanical decodings require interpretive choice: the written word “cane” exists in both English (walking stick) and Italian (dog), so choosing the right code already demands contextual judgment. Real interpretation involves encyclopedic knowledge—the vast, labyrinthine, rhizomatic network of interconnected cultural information that Eco contrasted with the reductive dictionary model of meaning. The encyclopedia is open-ended, multi-dimensional, and shot through with contextual, situational, and pragmatic information. It is the actual cognitive resource people draw on when they interpret signs, and it can never be fully formalized into neat code tables.

Connotation: how signs carry ideological freight

With codes established as the organizing systems within which signs operate, we can now examine what happens when signs begin to layer meanings on top of one another. This is Roland Barthes’s territory, and his analysis of denotation and connotation remains the single most useful practical tool in the semiotician’s kit.

The mechanism is precise and worth getting exactly right. Barthes, building on the Danish linguist Louis Hjelmslev’s formalization of Saussure, described two orders of signification. The first order is denotation: a signifier combines with a signified to produce a sign. A photograph of a rose denotes the flower. The word “car” denotes the vehicle. This is the literal, descriptive level—what Barthes called “the letter of the image”—requiring only basic perceptual or linguistic knowledge.

The second order is connotation, and its mechanism is the key insight. In connotation, the complete first-order sign—signifier and signified bound together—becomes the signifier of a new, second-order sign. An additional signified is attached to it. The photograph of the rose, now operating as a whole unit at the connotative level, signifies “romance” or “passion.” In Hjelmslev’s notation, where a basic sign is expressed as ERC (Expression–Relation–Content), connotation works by making the entire first system (E₁R₁C₁) the Expression of a second system: (E₁R₁C₁)R₂C₂. The first-order sign is, in Barthes’s vivid term, parasitized by the second-order meaning.

The signifiers of connotation Barthes called connotators, and the organized set of connotators in a given text constitutes its rhetoric. This is not rhetoric in the colloquial sense of empty persuasion, but rhetoric as the structured system of connotative signifiers at work in a specific communicative act.

Why connotation matters: the political dimension

Barthes saw connotation as inherently ideological, and his reasoning was structural, not merely political. Connotations are culturally specific—shaped by class, ethnicity, gender, generation, and the dominant value systems of a society. They are constructed, not natural. Yet they appear natural precisely because the first-order denotation provides an alibi. The literal image, the dictionary definition, the “obvious” meaning—these create a foundation of apparent objectivity that makes the layered connotative meanings seem self-evident, as if they arose from the thing itself rather than from cultural convention.

This is especially potent in photography. Barthes argued in “The Photographic Message” (1961) that the mechanical, analogical nature of the photograph makes it seem like a “message without a code”—a pure transcription of reality. But every photograph is worked up through framing, lighting, camera angle, focus, and selection. These operations “effectively belong to the plane of connotation,” yet the photograph’s appearance of objectivity disguises them. The denoted message, Barthes wrote, “naturalizes the system of the connoted message.” The literal image makes the ideological reading seem like a simple description of what is there.

The Panzani advertisement: connotation dissected

Barthes chose to demonstrate his method on a Panzani pasta advertisement precisely because advertising makes its signifying intentions explicit—the signs are “full, formed with a view to optimum reading.” The advertisement shows packets of pasta, a tin of concentrate, a sachet, some tomatoes, onions, peppers, and a mushroom, all emerging from a half-open string bag against a red background in yellows and greens.

Barthes identified three types of message operating simultaneously. The linguistic message includes the brand name “Panzani,” which denotes the company but—through its Italian-sounding vowel pattern—connotes Italianicity to a French reader. The coded iconic message (the connotative image) contains four discontinuous signs. First, the half-open bag with provisions spilling out connotes freshness and domestic preparation—a return from the market, implying hand-selected ingredients as opposed to mechanical mass-consumption. Second, the tomatoes, peppers, and tricolored hues (yellow, green, red) connote Italy, reinforcing the linguistic connotation. Third, the serried collection of diverse products together connotes a total culinary service—Panzani supplies everything needed for a balanced dish. Fourth, the overall composition evokes the tradition of the painted still life (nature morte), importing aesthetic prestige from art history.

Beneath all of this operates the non-coded iconic message—the literal denotation. Strip away all connotative readings, and what remains is identifiable objects in a spatial arrangement: tomato, bag, packet. This level requires only near-anthropological perceptual knowledge. But it is precisely this literal level that makes all the connotative work possible, because it provides the naturalistic foundation: the connotations appear to emerge from the objects themselves, as if Italianicity were a property of tomatoes rather than a French cultural construction.

Barthes noted that the connotation of Italianicity was “specifically French, based on a familiarity with certain tourist stereotypes”—an Italian would barely perceive it. This observation reveals a crucial property of connotation: it is code-dependent. Without the relevant cultural code, the connotation does not fire. This is why the same image can mean radically different things in different cultural contexts, and why connotative analysis must always specify the codes it assumes.

How connotations stabilize and naturalize

Connotations become naturalized—transformed from cultural constructions into apparent common sense—through several interlocking mechanisms. Repetition across cultural forms is the most powerful: when the same connotative meanings are reproduced across advertising, film, news, education, and everyday conversation, they solidify into taken-for-granted truths. Photography’s special status accelerates this process by providing an apparently objective ground for subjective cultural meanings. And over time, connotations that have achieved sufficient dominance begin to look like denotations—they become so habitual that they appear to be simply what the sign means, rather than one culturally loaded reading among many.

Myth: how culture disguises itself as nature

Barthes’s theory of myth takes the machinery of connotation and reveals its deepest ideological operation. In “Myth Today,” the theoretical essay concluding Mythologies (1957), Barthes defined myth not as classical fable but as “a type of speech,” “a system of communication,” “a mode of signification, a form.” Myth is not defined by its content—any subject can be mythologized—but by the way it processes meaning.

The structural mechanism is identical to connotation: the sign of the first-order system becomes the signifier of a second-order system. But Barthes introduced specific terminology to track what happens at each level. The first-order sign, which was a full, rich meaning—a complete sign with its own history, context, and specificity—becomes what Barthes called the form in the mythical system. It is simultaneously still a first-order meaning (it was full, experienced, specific) and now a second-order form (emptied, impoverished, held at a distance). The new signified that myth attaches to this form Barthes called the concept. And their union—the mythical sign—he called the signification.

The dual status of the form is the engine of myth. Barthes was emphatic: “The form does not suppress the meaning, it only impoverishes it, it puts it at a distance, it holds it at one’s disposal.” The first-order meaning is not destroyed—myth needs it. But it is drained of its specific historical content and made available as raw material for the concept to fill. The meaning “loses its value, but keeps its life, from which the form of the myth will draw its nourishment.” This produces what Barthes described as “a constant game of hide-and-seek between the meaning and the form”—the defining oscillation of mythological signification.

The Paris Match cover: myth at work

Barthes’s central example remains devastating in its clarity. At the barber shop, he picks up a copy of Paris-Match. On the cover, a young Black soldier in French military uniform gives the French salute, eyes lifted toward what is probably the tricolor flag.

At the first order, this is a full sign: a specific person with a biography, a history, a particular moment. But myth takes this sign and makes it a form. “One must put the biography of the Negro in parentheses,” Barthes wrote, “if one wants to free the picture, and prepare it to receive its signified.” The soldier’s specific identity is drained; he becomes available as raw material. The concept that fills this form is “a purposeful mixture of Frenchness and militariness”—what Barthes called French imperiality. The signification that results is the claim that France is a great empire, that all her sons serve faithfully under the flag without racial discrimination, and that there is no better refutation of critics of colonialism than this willing Black soldier.

The historical context makes the ideological operation vivid. This was published in July 1955—just over a year after France’s crushing defeat at Dien Bien Phu, and during the escalating Algerian War of Independence. The myth does not deny French imperialism; on the contrary, it needs to actualize its presence. What myth eliminates is the contingent, historical, fabricated quality of colonialism—the fact that it is a political project with opponents, victims, and alternatives. Myth transforms this contested political reality into what appears to be a natural fact: of course all French citizens serve their nation; of course the empire is harmonious.

Barthes captured the core operation in his most famous formulation: “Myth transforms history into nature.” Complex socio-historical realities—contingent, fabricated, contested, politically motivated—are presented as self-evident facts, as eternal truths, as natural states of affairs. “Everything happens as if the picture naturally conjured up the concept, as if the signifier gave a foundation to the signified.” The intention behind the myth is “frozen into something natural; it is not read as a motive but as a reason.”

The depoliticizing function follows directly. Barthes stated: “Myth is depoliticised speech,” where “political” encompasses all human relations in their actual structure as power to transform the world. Myth strips phenomena of their historical context and contradictions, converting usage (doing) into nature (being), presenting culturally specific ideologies as universal and timeless. “Myth does not deny things,” Barthes wrote. “Its function is to talk about them; simply, it purifies them, it makes them innocent, it gives them a natural and eternal justification, it gives them a clarity which is not that of an explanation but that of a statement of fact.”

Other mythological operations

Barthes applied this machinery across French culture with relentless analytical energy. Professional wrestling, he argued, functions as a mythological system where every gesture is a perfectly transparent sign: the villain’s physique embodies evil, the hero’s victory enacts justice, and the spectacle mythologizes moral clarity—“the pure gesture which separates Good from Evil”—abolishing the real-world ambiguity of ethical judgment. Wine in French culture operates as myth by signifying health, strength, community, and French identity itself, naturalizing heavy drinking as an inherent quality of Frenchness while obscuring alcohol’s dangers and the economics of wine production, including colonial labor exploitation. In each case, the analytical move is the same: identify the first-order sign, show how it is emptied into form, name the ideological concept that fills it, and reveal how the resulting signification transforms a historical construction into an apparent natural fact.

The post-structuralist challenge to mythological stability

Barthes’s myth analysis assumes that while myth disguises meanings as natural, the analyst can unmask them—can identify the concept hiding behind the form, can restore the history that myth suppresses. But Jacques Derrida’s post-structuralist critique challenges this confidence in ways that are both troubling and productive.

Derrida’s concept of différance—a deliberate misspelling that fuses “to differ” (spatial dimension: signs get meaning through difference from other signs) with “to defer” (temporal dimension: meaning is perpetually postponed as each sign refers to further signs)—radicalizes the Saussurean insight that launched structuralism. If meaning is purely differential, with no positive content to any sign, then no sign can ever achieve the stable, self-present meaning that Barthes’s analysis seems to require. Each signifier leads to a signified that immediately functions as another signifier, in an endless chain. The “transcendental signified”—the ultimate foundational meaning that would anchor all other signs—never arrives.

Derrida’s critique does not destroy Barthesian myth analysis so much as complicate it. The mythologist who claims to unmask French imperiality behind the Paris Match cover is herself producing signs about signs, and her analysis is subject to the same deferral and differential instability she diagnoses in myth. There is no position outside the play of signifiers from which to deliver a final, definitive reading. Derrida’s famous formulation “il n’y a pas de hors-texte” (there is no outside-text) does not mean reality does not exist—it means there is no access to reality that is not already mediated by signification. We cannot step outside the web of signs to reach a pristine, pre-linguistic truth against which myth could be definitively measured.

This challenge is best understood not as a refutation but as a productive tension. Barthesian myth analysis retains enormous practical power—it reveals ideological operations that would otherwise remain invisible. But Derrida’s critique reminds us that the analyst’s own categories are historically situated, that codes are never perfectly sealed, and that every act of demystification produces new texts open to further interpretation. The play of signifiers means that no single reading, however rigorous, can claim to be the final one.

Ideology and power: how signs serve domination

Myth, in Barthes’s account, is the mechanism by which culture disguises itself as nature. But Barthes was working a specific seam of a much broader vein: the relationship between sign systems and structures of power. To understand how signs serve ideological functions, we need to examine the specific mechanisms through which meaning upholds—or challenges—relations of domination.

The machinery of ideological signification

John B. Thompson, in Ideology and Modern Culture (1990), provided the most systematic taxonomy of how ideology operates through semiotic construction. He defined ideology as “the ways in which meaning serves to establish and sustain relations of domination,” and identified five principal modes.

Naturalization makes socially constructed arrangements appear to be natural facts or inevitable outcomes of nature itself. Gender roles presented as biologically determined rather than historically produced; the nuclear family presented as the “natural” form of human organization; market capitalism presented as reflecting inherent human nature—all are naturalization operations. The semiotic mechanism is precisely Barthes’s myth: history transformed into nature.

Reification makes abstract social relations appear thing-like, converting transitory processes into permanent objects. The phrase “the market demands austerity” treats an abstract system of social relations as a concrete agent with its own volition. Linguistically, reification operates through nominalization (converting verbs into nouns: “management fired workers” becomes “downsizing occurred”) and passivization (removing agents: “wages were cut” obscures who cut them). These grammatical operations are semiotic operations: they restructure the sign system to conceal agency and present human choices as impersonal forces.

Universalization presents particular interests as universal ones. Tax cuts primarily benefiting the wealthy are framed as “good for the economy”—as if the economy were everyone’s economy equally. Free trade agreements negotiated to benefit specific industries are presented as beneficial for all nations. The mechanism is the erasure of differential impact: signs are constructed to obscure that particular groups benefit and others lose.

Euphemization substitutes signs with positive connotations for signs that would reveal harsh realities. “Collateral damage” for civilian deaths. “Downsizing” for mass layoffs. “Enhanced interrogation” for torture. “Developing nations” for economically exploited countries. Each substitution is a precise semiotic operation: replacing one signifier with another to shift the connotative field while maintaining the same denotative reference. The mechanism works because it exploits the gap between denotation and connotation—the euphemism denotes roughly the same thing but connotes something entirely different.

Legitimation represents relations of domination as legitimate through chains of reasoning (rationalization), appeals to universal benefit, and embedding in narratives of cherished tradition. The semiotic dimension is the construction of narrative frameworks that make existing power arrangements appear justified, reasonable, and deeply rooted.

Gramsci, hegemony, and the semiotics of common sense

Antonio Gramsci’s concept of hegemony provides the broader political framework within which these mechanisms operate. For Gramsci, a dominant class maintains power not primarily through force but through ideological and cultural leadership—securing the “active or passive consent” of subordinate classes by making the dominant worldview appear to be simple common sense.

Common sense, in Gramscian terms, is the incoherent but widely accepted blend of ideas, assumptions, and beliefs that sustain hegemonic norms. It functions semiotically through the naturalization of discourse: dominant cultural values become so thoroughly embedded in the codes of everyday communication that they seem self-evident, requiring no justification. The ruling class, Gramsci argued, “strives to impart a supra-class, eternal character to the ideological sign”—exactly the operation Barthes would later describe as myth. Hegemony is maintained not by controlling what people think but by controlling the codes within which thinking occurs: the categories, frameworks, and “obvious” assumptions that structure what it is even possible to think.

Valentin Voloshinov, in Marxism and the Philosophy of Language (1929), offered an even more radical formulation: “The sign becomes an arena of class struggle.” Because different classes use the same language, “differently oriented accents intersect in every ideological sign”—the same sign carries competing social evaluations, can mean different things in different class positions. The ruling class works to make signs uni-accentual: to fix them with a single, apparently natural meaning, suppressing the competing accents. But Voloshinov argued that this fixing is never complete. In times of social crisis, the suppressed accents resurface, and “any current truth must inevitably sound to other people as the greatest lie.” Signs that lose all contestation become dead—“not of live social intelligibility but of philological comprehension.” The life of signs, for Voloshinov, is their contested, multi-accentual character.

Hall’s encoding/decoding: the asymmetry of communication

Stuart Hall’s encoding/decoding model (1973/1980) translates these insights about ideology and signs into a working framework for analyzing mass communication. Hall’s key innovation was to show that encoding and decoding are not symmetrical processes—the meaning a producer puts into a text is not necessarily the meaning an audience takes out.

The mechanism works as follows. Encoding is the process by which producers construct messages within sign systems. A historical event cannot be transmitted raw—it must be translated into the discursive forms of the medium. A television newscast must turn events into stories, using professional codes (framing, narrative structure, interview conventions), technical codes (camera angles, editing, graphics), and ideological codes (assumptions about what matters, who deserves attention, what constitutes reasonable opinion). These codes are not neutral: they embed what Hall called a preferred meaning—the reading toward which the text’s structure of signification pushes the audience.

Decoding is the audience’s active production of meaning from the encoded text. Hall insisted this is not passive reception. The viewer brings their own framework of knowledge, class position, cultural background, and ideological codes to the interpretive act. Because encoding and decoding are “over-determined, relatively autonomous” moments, each with its own conditions of existence, “no one moment can fully guarantee the next moment with which it is articulated.” The degree of understanding depends on the degree of overlap between the codes of encoding and decoding—and this overlap is never guaranteed.

Hall identified three hypothetical decoding positions. In the dominant-hegemonic position, the viewer operates within the dominant code, accepting the preferred reading as natural and self-evident. A news report on economic austerity is understood as presenting necessary sacrifice for national well-being. In the negotiated position, the viewer accepts the dominant reading at an abstract level but modifies it based on local experience: yes, austerity may be nationally necessary, but my sector should be exempt because of our specific conditions. In the oppositional position, the viewer understands the dominant code—grasps the preferred reading—but decodes within an alternative framework, rejecting the dominant interpretation entirely: every mention of “national interest” is re-read as “class interest.”

These positions exist on a continuum rather than as fixed points, and they are ideological in nature—they refer to social groups and their relationship to dominant meaning systems, not to individual psychological quirks. Hall’s model fundamentally challenged the transmission model of communication (sender → message → receiver) by showing that meaning is produced at both ends of the process, that the text is polysemic (open to multiple readings) though not equally open, and that the struggle over meaning is a struggle over power.

The materialist challenge: signs versus material conditions

Hall’s framework, and the Barthesian apparatus it builds on, operate largely within the domain of signification. But a persistent and important critique, rooted in Marxist materialism, asks whether this focus on signs risks obscuring the material conditions that ultimately determine social life.

The materialist argument, in its strongest form, holds that semiotics is idealist—that it treats signs and codes as autonomous systems divorced from the economic base of society. Voloshinov himself insisted that “the laws of this reality are the laws of semiotic communication and are directly determined by the total aggregate of social and economic laws.” Signs are not free-floating; they are produced under specific material conditions, by people in specific class positions, using materially produced technologies. A semiotic analysis that examines only the internal structure of signs without asking who produced them, under what conditions, and in whose interest, risks reproducing the very mystification it claims to unmask.

Raymond Williams sharpened this critique with his concept of structures of feeling—the lived quality of experience in a particular time and place that exists in the gap between what official sign systems articulate and what people actually feel. Williams argued that there is always a dimension of experience that has not yet been codified into formal signs—it is pre-emergent, not yet fully articulated in ideological or semiotic terms. This challenges any semiotic analysis that treats codes as exhaustive: there is always a residue of lived experience that escapes the sign system, and this residue is often precisely where historical change originates.

The most productive response to the materialist critique is not to abandon semiotic analysis but to situate it. Signs do not merely represent material relations—they actively participate in producing, maintaining, and sometimes challenging them. Marx’s concept of commodity fetishism is itself a semiotic insight: the commodity form is a sign operation in which social relations between people appear as relations between things. The price tag, the brand, the advertising image—all replace and obscure the social relations of labor. Semiotic analysis and materialist analysis are not alternatives but complementary: the first reveals how meaning is constructed; the second reveals whose interests that construction serves and what material conditions make it possible.

Complementary frameworks: expanding the analytical toolkit

Barthes and Hall gave semiotics its ideological cutting edge. But several other frameworks have significantly expanded what semiotic analysis can do, each addressing limitations in the foundational models.

Greimas’s narrative semiotics: the deep structure of stories

Algirdas Julien Greimas addressed a problem that neither Saussure nor Barthes fully solved: how meaning is organized across extended texts, particularly narratives. Where Saussure analyzed the sign and Barthes analyzed the layering of meaning, Greimas sought the deep syntactic structures that underlie all narrative action.

His actantial model, developed in Sémantique structurale (1966), breaks any action into six functional roles organized on three axes. The axis of desire connects the Subject (the entity directed toward a goal) to the Object (the goal itself). The axis of power connects the Helper (what assists the quest) to the Opponent (what obstructs it). The axis of transmission connects the Sender (what instigates the quest) to the Receiver (who benefits from its completion). These are not concrete characters but abstract structural positions: a single character can fill multiple actant roles simultaneously, and a single role can be distributed across multiple characters or even embodied by abstract forces.

The model’s analytical power lies in what it reveals about the deep structure of any narrative—including narratives that do not look like stories on their surface. Apply the actantial model to a political speech, and the Subject might be the nation, the Object prosperity, the Sender the people’s mandate, the Receiver the citizenry, the Helper the proposed policy, and the Opponent the rival party. Apply it to an advertisement, and the Subject is the consumer, the Object is satisfaction or social success, the Sender is desire (activated by the ad), the Receiver is the consumer again, the Helper is the product, and the Opponent is the problem the product solves. The model does not tell you what a narrative means, but it reveals the structural logic of agency, desire, and conflict that organizes meaning.

Greimas also developed the semiotic square, which generates analytical positions from simple binary oppositions. Starting from two contrary terms (say, life and death), the square produces their contradictory terms (non-life and non-death), yielding four positions connected by three types of relationship: contrariety (life vs. death: both cannot be simultaneously true, but both can be simultaneously false), contradiction (life vs. non-life: strict either/or), and implication (non-death implies life). The square’s analytical value comes from the compound positions it generates: the complex term where both contraries are present (life AND death: the undead, the martyr), and the neutral term where neither is present (neither life nor death: the unborn, the transcendent). Applied to concepts, the square maps ideological terrain with remarkable precision, revealing positions that a simple binary would obscure.

His canonical narrative schema further organized narrative into four phases—manipulation (a sender motivates the subject), competence (the subject acquires the ability to act), performance (the subject performs the decisive action), and sanction (the action is evaluated and rewarded or punished)—providing a grammar of narrative that analysts can apply across genres, media, and cultures.

Lotman’s semiosphere: signs in ecological context

Yuri Lotman, founder of the Tartu-Moscow Semiotic School, introduced the concept of the semiosphere in 1984 to challenge the atomistic tendency of semiotic analysis—the habit of studying individual signs or texts in isolation. Lotman argued, by analogy with Vernadsky’s biosphere, that “the unit of semiosis, the smallest functioning mechanism, is not a separate language but the whole semiotic space of the culture in question.”

The semiosphere is the total semiotic environment within which all sign systems operate. It precedes and is presupposed by any specific system of signs: “Without the semiosphere, language not only does not function, it does not exist.” Its internal space is heterogeneous, composed of diverse, incommensurable semiotic systems—verbal language, visual codes, musical notation, gestural systems, architectural conventions—that cannot be adequately translated into one another. This asymmetry is not a defect but a generative force: because translation between systems is always imperfect, every attempt at translation produces new information not present in the original. For Lotman, “to think means to translate from one language into another,” making thought itself the most powerful generator of new meaning.

The boundary of the semiosphere is its most important structural feature—a “bilingual mechanism” that translates external communications into the internal languages of the culture and vice versa, functioning like sensory receptors that convert stimuli into neural signals. The internal dynamics of the semiosphere are organized around a center-periphery tension: the center contains the most developed, normalized, canonical codes; the periphery is “the area of accelerated semiotic processes” where new texts are generated, codes are less rigid, and innovation occurs. Periodically, what was peripheral moves to the center and vice versa—a dynamic that Lotman used to explain cultural revolutions and paradigm shifts.

Social semiotics and multimodality: signs in the real world

While Greimas sought deep structures and Lotman mapped semiotic ecologies, social semiotics turned attention to how real people make meaning in concrete social situations. Developed by Robert Hodge and Gunther Kress in Social Semiotics (1988) and extended by Kress and Theo van Leeuwen in Reading Images (1996), this approach makes a fundamental break with structuralist semiotics on the question of the sign’s nature.

Classical Saussurean semiotics treats the sign as arbitrary: the relationship between signifier and signified is conventional, not motivated. Social semiotics argues instead that signs are motivated: sign-makers, driven by their interests (a complex condensation of their social, cultural, and psychological history), choose the most apt form for the meaning they want to express. The sign is not selected from a pre-existing code but actively designed in the moment of communication.

This shift has several consequences. Social semiotics replaces the concept of codes (fixed, pre-existing rule systems) with semiotic resources—the actions, materials, and artifacts available for communicative purposes, together with the ways they can be organized. Resources have a meaning potential based on their cultural history and a set of affordances based on their material properties and social uses. Spoken language affords temporal sequencing; images afford spatial simultaneity. Color affords emotional expression in ways that mathematical notation does not. Each mode—language, image, sound, gesture, spatial arrangement—has its own logic, its own affordances, and its own limitations.

Drawing on Michael Halliday’s systemic functional linguistics, Kress and van Leeuwen argued that all semiotic modes must serve three simultaneous metafunctions: the ideational (representing the world), the interpersonal (enacting social relationships), and the textual (organizing coherent communication). They developed a “grammar of visual design” that analyzes images through representational meanings (what is depicted and how—through narrative vectors or conceptual classifications), interactive meanings (how the image relates to the viewer—through gaze contact, social distance, camera angle), and compositional meanings (how elements are arranged—information value of left/right, top/bottom, center/margin, plus salience and framing).

Multimodality—the study of how multiple semiotic modes combine in communication—is social semiotics’ most important contribution to contemporary analysis. In the digital age, meaning is almost never made through a single mode. A website combines text, image, color, layout, animation, and sound. A social media post integrates photographs, captions, hashtags, emojis, and platform-specific conventions. Understanding how these modes interact—whether they reinforce, complement, or contradict each other—requires analytical tools that go beyond the sentence-level analysis of traditional linguistics and the single-image analysis of classical visual semiotics.

The phenomenological challenge: meaning before signs

Each of these frameworks—Greimas’s narrative grammar, Lotman’s semiosphere, Kress and van Leeuwen’s social semiotics—expands the reach of semiotic analysis. But a fundamental philosophical challenge, rooted in phenomenology, asks whether the semiotic project has limits that no expansion can overcome.

Maurice Merleau-Ponty, in Phenomenology of Perception (1945), argued that before any sign system operates, we are already meaningfully engaged with the world through our bodies. We do not first perceive a neutral object and then apply the sign “chair” to it. We immediately apprehend the chair as something-to-sit-on—its meaning is inseparable from our bodily capacities, our motor habits, our history of physical engagement with similar objects. Merleau-Ponty replaced Descartes’s “I think” with a tacit cogito: “I can”—bodily capability precedes intellectual cognition. The competent typist does not consciously process key positions through a sign system; the knowledge is “in the hand.”

This challenges the structuralist claim that meaning is purely differential—that signs have no positive content, only relations to other signs. Merleau-Ponty countered that meaning is grounded in embodied experience: mountains are tall “for us” because of our bodily scale; spaces are passable or impassable relative to our physical capacities. The world shows up as meaningful not because of a system of differential signs but because of our bodily engagement with it. There is, Merleau-Ponty insisted, a layer of pre-predicative meaning—meaning that exists before any sign is applied—that cannot be fully captured by semiotic analysis.

This does not refute semiotics, but it establishes a limit. Semiotic analysis operates on meaning that has already been articulated in sign systems. But there is always a residue of embodied, pre-linguistic experience that feeds into semiotic systems without being reducible to them. Language itself, Merleau-Ponty argued, is not a transparent system of arbitrary signs but is rooted in bodily expression—each language is “a distinct and ultimately untranslatable manner of singing the world.” The phenomenological critique is best taken not as a rejection of semiotics but as a reminder that signs emerge from, and always remain connected to, a more fundamental layer of embodied being-in-the-world. Any semiotic analysis that forgets this ground risks mistaking the map for the territory—treating the sign system as if it were the totality of meaning, when it is in fact a culturally constructed layer floating on an ocean of embodied experience.

Specialized domains: where semiotic analysis meets specific media

The frameworks developed so far—codes, connotation, myth, ideology, narrative structure, the semiosphere, multimodality—are general-purpose tools. But specific media present specific analytical challenges, and semioticians have developed specialized approaches for each. Here we preview the major domains; the following chapter will explore them in depth.

Visual semiotics: Panofsky’s enduring framework

Erwin Panofsky, working in art history before semiotics existed as a formal discipline, developed a three-level interpretive scheme in Studies in Iconology (1939) that maps remarkably well onto semiotic categories. His pre-iconographic level identifies pure forms as representations of natural objects and events—a man, a table, a gesture of grief. This corresponds roughly to denotation. His iconographic level connects artistic motifs with cultural themes and stories—thirteen men at a table in a specific arrangement depicts the Last Supper. This requires knowledge of cultural conventions and corresponds to connotation. His iconological level seeks the deep symbolical values that a work expresses—“the basic attitude of a nation, a period, a class, a religious or philosophical persuasion”—often communicating things the creator did not consciously intend. This corresponds to Barthes’s myth or ideology: the level at which cultural values are naturalized and a work becomes a symptom of its era.

Visual signs differ from linguistic signs in ways that create distinctive analytical challenges. They tend to be more motivated (a portrait resembles its subject; a photograph has a causal indexical connection to what it depicts) and more analog (they involve continuous gradations of color, brightness, and form rather than discrete either/or units). This means visual signs resist the clear articulation into minimal units that Saussure identified in language, and it partly explains why efforts to construct a visual “grammar” have been more contested than linguistic grammar.

Film semiotics: Metz and the language of cinema

Christian Metz’s contribution to film semiotics centered on a deceptively simple question: is film a language? His answer, developed across Film Language (1968) and Language and Cinema (1971), was precise and productive. Film is a langage (a language-like communicative system) but not a langue (it lacks a strict grammar equivalent to natural language). Film’s basic unit, the shot, is iconic rather than arbitrary—it resembles what it depicts. Film has no double articulation: there are no meaningless “film phonemes” that combine into meaningful units the way phonemes combine into morphemes. And film has no fixed vocabulary—each shot is essentially unique, depicting whatever the camera was pointed at.

Metz’s grande syntagmatique attempted to catalogue the segment types that filmmakers use to construct meaning—from the autonomous shot through various forms of parallel, alternating, and sequential editing to the scene and the ordinary sequence. His taxonomy identified eight types of autonomous segments, providing a vocabulary for analyzing how films organize time, space, and causality through editing. More important than the specific taxonomy was the insight that cinema deploys multiple overlapping codes simultaneously—codes of lighting, framing, editing, sound, narrative structure, and cultural convention all operate in the same moment of film, requiring the viewer to process multiple semiotic systems at once.

Broader horizons: fashion, architecture, digital media

Barthes himself extended semiotic analysis to fashion in The Fashion System (1967), analyzing how fashion magazines construct meaning through the “written garment”—the verbal description of clothing as a sign system with its own signifiers (garment features) and signifieds (social meanings). Eco analyzed architecture as a sign system where buildings both function and communicate: a staircase denotes the ability to go up and connotes specific cultural values through its style. And the digital environment has created entirely new semiotic challenges: interactivity makes the user a co-producer of meaning rather than a passive decoder; hypertextuality breaks the linear sequence that traditional semiotic models assume; and the collapse of the producer/consumer distinction on social media platforms challenges the very encoding/decoding framework Hall developed for broadcast media.

Each of these domains reveals that while the foundational semiotic concepts—sign, code, denotation, connotation, myth—travel remarkably well across media, each medium’s specific material properties (its affordances, in social-semiotic terms) create distinctive meaning-making possibilities and constraints. The next chapter will explore these specialized domains in depth, examining how semiotic analysis adapts when it encounters the particular challenges of visual composition, moving images, spatial design, embodied performance, and digital interaction.

5 · Specialized Domains

The semiotic frameworks developed so far—Saussure’s relational value, Peirce’s triadic sign, Barthes’s connotation and myth, Eco’s code theory, Greimas’s structural semantics, Lotman’s semiosphere—were built to be general. They claim to describe how meaning works regardless of medium. The real test of that claim comes when we leave the comfortable territory of verbal language and apply these tools to images, buildings, digital interfaces, living organisms, and music. Each domain resists semiotic analysis in its own characteristic way, and each resistance teaches us something about both the domain and the limits of the apparatus.

5.1 · Visual semiotics: how images organize meaning

Panofsky’s three-level iconographic method, introduced earlier, gives us a vertical axis for reading images—from pre-iconographic description through iconographic identification to iconological interpretation. But images also organize meaning horizontally, through their internal compositional grammar. The most systematic account of this grammar comes from Gunther Kress and Theo van Leeuwen’s Reading Images: The Grammar of Visual Design (1996), which translates Halliday’s systemic functional linguistics into visual terms.

Compositional grammar. Kress and van Leeuwen argue that the placement of elements within a frame is not arbitrary but follows culturally coded spatial logic. In cultures that read left-to-right, the left position codes the Given—information treated as already known, established, or taken for granted—while the right position codes the New—information presented as novel, at issue, or noteworthy. A magazine advertisement that places a “before” image on the left and an “after” image on the right is exploiting this Given/New structure: the problem is familiar; the solution is news. The vertical axis works differently: the top position codes the Ideal—the generalized, aspirational, or emotionally appealing essence—while the bottom codes the Real—the specific, practical, grounded, or informational. An advertisement with a glamorous model at the top and product specifications at the bottom literalizes this structure: the promise floats above the evidence. Finally, in centered compositions, the center functions as the nucleus and surrounding elements as margins—dependent, subordinate, or supplementary. These are tendencies, not iron laws; they describe a semiotic resource, not a universal constraint. But their predictive power across Western visual media is remarkably consistent.

Gaze and power. The relationship between depicted subjects and viewers is itself a sign system. When a depicted person looks directly at the viewer, Kress and van Leeuwen call this a “demand” image: the subject demands something from the viewer—recognition, engagement, empathy, identification. When the subject looks away, unaware of the viewer’s gaze, it is an “offer” image: the subject is offered up for inspection, made available as an object of contemplation. The demand/offer distinction intersects with social distance (close-up implies intimacy; long shot implies detachment) and angle of interaction (low angle grants depicted subjects power over the viewer; high angle reverses that relation; frontal angle implies involvement; oblique angle implies detachment).

Laura Mulvey’s concept of the male gaze, introduced in “Visual Pleasure and Narrative Cinema” (1975), adds a gendered dimension. Mulvey argued that classical Hollywood cinema structures looking around three aligned gazes—the camera’s gaze, the characters’ gazes within the narrative, and the spectator’s gaze—and that all three are organized around a heterosexual male perspective. Women are coded with what Mulvey calls “to-be-looked-at-ness”: they function as visual spectacle while male characters drive narrative action. The spectator identifies with the active male protagonist, adopting his gaze as their own. The mechanism operates through what Freud called scopophilia—pleasure in looking—split into voyeuristic pleasure (watching without being seen) and narcissistic identification (recognizing oneself in an idealized image). Mulvey’s framework has been challenged—bell hooks argued it ignored race and Black female spectatorship; Gaylyn Studlar proposed masochistic rather than sadistic identification—but it remains indispensable for understanding how visual texts encode power through the structure of looking itself.

Color as semiotic mode. Color is not decorative but meaning-bearing. Every culture maintains color codes—red for danger or passion, white for purity or mourning depending on cultural context—but semiotic analysis goes beyond cataloguing associations. Kress and van Leeuwen identified color saturation, brightness, and modulation as modality markers: indicators of how “real” or “true” a visual representation claims to be. The key insight is that modality is context-dependent. In a naturalistic coding orientation (typical of documentary or journalism), moderate saturation matches how the eye sees the world, so it signals high modality—truthfulness. Hyper-saturated images drop in naturalistic modality: they look “too vivid to be real.” But in a sensory coding orientation (advertising, fashion, art), high saturation signals high modality—it conveys the “more-than-real” intensification that these genres treat as truth. A desaturated news photograph and a hyper-saturated perfume advertisement can both claim to be showing us something real, but they operate within different coding orientations, and reading them requires recognizing which orientation is in play.

Visual metaphor. Charles Forceville’s work on pictorial metaphor demonstrates that metaphors are not confined to language. A visual metaphor works by mapping attributes of one domain onto another through pictorial means—a hybrid metaphor fuses two referents into a single impossible image (a running shoe with cheetah spots), while a contextual metaphor replaces one referent with another in an unexpected setting. The mechanism is the same as verbal metaphor—structured mapping between source and target domains, as Lakoff and Johnson described—but the visual mode offers resources unavailable to language: simultaneity (both domains visible at once), spatial relations, and the immediate perceptual impact of visual resemblance.

Worked example: analyzing a magazine cover. Consider a hypothetical fitness magazine cover. A muscular figure occupies the center (nucleus of composition), facing the camera with direct eye contact (demand image—demanding admiration, aspiration, or identification). The figure is photographed from a slightly low angle (granting the subject power). High color saturation and sharp focus operate within a sensory coding orientation, producing high modality—this body is presented as more real than real, the aspirational truth the magazine sells. The magazine title floats at the top (Ideal zone—the brand promise), while cover lines listing specific workout routines and diet plans cluster at the bottom (Real zone—the practical deliverables). Left-side text references familiar fitness concepts (Given), while the right-side announces “NEW: 30-Day Transformation Plan” (New). The male gaze may structure the image even when the depicted subject is male—the body is presented for scopophilic consumption, its musculature coded with to-be-looked-at-ness typically reserved for female subjects in Mulvey’s analysis, suggesting how fitness culture has extended the logic of visual objectification across genders. The entire cover is a mythological operation in Barthes’s sense: it takes culturally specific, commercially motivated body ideals and presents them as natural truths about health and human potential.

5.2 · Digital semiotics: platforms, memes, and the crisis of indexicality

Digital environments constitute an entirely new semiosphere, one in which the interface itself is a sign system, the boundaries between producer and consumer dissolve, and the foundational indexicality of photographic media is under existential threat.

The interface as sign system. Every element of a graphical user interface is a sign. The desktop metaphor, developed at Xerox PARC and popularized by Apple’s 1984 Macintosh, organized the computer screen around icons that resembled office objects—files, folders, a trash can, a clipboard. These were skeuomorphic signs: digital objects that retained the visual appearance of physical predecessors. In Peircean terms, they were predominantly iconic—signifying through resemblance. The transition from skeuomorphism to flat design (beginning around 2012–2013 with Microsoft’s Metro and Apple’s iOS 7) marks a semiotic shift from iconic to symbolic signification: abstract geometric shapes whose meanings are learned through convention rather than recognized through resemblance. A skeuomorphic leather-textured notepad icon looks like a notepad. A flat colored rectangle labeled “Notes” signifies through arbitrary convention. This transition itself is culturally diagnostic: it became possible only when a generation of users had internalized digital conventions deeply enough that they no longer needed physical-world anchors.

Emoji as semiotic hybrids. Emoji complicate Peirce’s taxonomy because they operate across all three sign modes simultaneously. A smiley face emoji is iconic (it resembles a smiling face), symbolic (its specific meaning—friendliness, irony, passive aggression—depends on contextual convention), and indexical (it points to the sender’s emotional state, functioning as a trace of affect). The cultural variability of emoji interpretation reveals their symbolic dimension most clearly: the 🙏 emoji (Unicode: “Person with Folded Hands”) is read as prayer in Western Christian contexts, as gratitude or itadakimasu in Japanese contexts, and as namaste in South Asian contexts. The same visual form, identical at the iconic level, carries divergent meanings because different cultural codes are applied to decode it. Emoji are not the universal pictographic language their enthusiasts sometimes claim; they are culturally embedded signs that appear transparent precisely because their iconic dimension masks their conventional, symbolic operation.

Memes as distributed semiosis. Internet memes represent something genuinely new in the history of sign systems: meaning produced through collective, iterative transformation across thousands of participants. Limor Shifman defines memes not as single viral units but as groups of digital items sharing common characteristics of content, form, and stance, created with awareness of each other, and circulated through imitation and transformation. The mechanism is essentially intertextual: each meme iteration derives its meaning from its relationship to all other iterations. A meme template is a syntagmatic structure with paradigmatic slots—the “Distracted Boyfriend” image provides a fixed spatial and narrative syntax (man looks away from partner toward another woman), while different iterations fill those slots with different referents (old technology/user/new technology; obligation/person/temptation). Each new fill is a paradigmatic substitution within a stable syntagm, exactly the operation Saussure described—but now performed collaboratively, at scale, and at speed. The meme’s meaning is not located in any single iteration but distributed across the entire set. This is semiosis without a singular author, a concrete demonstration of Barthes’s insight that the text is a tissue of quotations from innumerable centers of culture.

Platform affordances as meaning constraints. The platform on which a sign circulates is not a neutral channel but a system of constraints that shapes what can be meant. Twitter’s character limits (originally 140, now 280) enforce aphoristic compression, favoring wit, polemic, and phatic communion over sustained argument. Instagram’s visual emphasis and filter infrastructure produce what Kress and van Leeuwen would recognize as a sensory coding orientation—high saturation, careful composition, aestheticized life. TikTok’s short-form video format and algorithmic “For You” page create a semiotic environment dominated by rhythm, embodied performance, and extreme intertextuality (sounds are shared across millions of videos, each new use a citation of previous uses). These are not mere technical constraints but semiotic affordances: they determine which sign functions can be dominant. Twitter privileges the referential and conative; Instagram privileges the poetic (visual form for its own sake); TikTok privileges the phatic and the poetic simultaneously.

Algorithmic meaning-making. When a recommendation algorithm selects and sequences content for a user, it performs what are fundamentally editorial and authorial operations—choosing what to include, what to exclude, and in what order elements appear. These are the operations of syntagmatic construction (sequencing) and paradigmatic selection (choosing from alternatives). The algorithm is, in semiotic terms, an author, even though it has no intentions, no consciousness, and no interpretant in Peirce’s sense. This creates a genuine theoretical puzzle: Peirce defined semiosis as involving a sign, an object, and an interpretant (the effect on some mind). If the algorithm produces signs without an interpretive mind, is this semiosis? Or is the algorithm merely a mechanism that produces sign-vehicles whose semiosis occurs only when a human user encounters them? The question matters because algorithmic curation does not merely present pre-existing meaning—it creates new meaning through juxtaposition, emphasis, and omission.

Deepfakes and the crisis of indexicality. Photographs have functioned as evidence for over a century because of their indexical character: a photograph is caused by light reflecting off actual objects and inscribing a trace on film or sensor. This causal connection is what Peirce meant by indexicality—the sign bears a physical, existential relation to its object, like a fingerprint or a weathervane. Barthes, in Camera Lucida, grounded the photograph’s affective power in this indexical trace: the knowledge that this light actually touched this person. Deepfakes sever this causal connection entirely. A deepfake video is generated by neural networks extrapolating from training data; no light from the depicted scene ever touched any sensor. The image is iconic (it resembles reality) and potentially symbolic (it can convey narrative), but it is not indexical—it bears no causal trace of what it depicts. The result is an epistemic crisis that operates at two levels. First, specific deepfakes can deceive. Second, and more insidiously, the mere possibility of deepfakes degrades trust in all photographic evidence—the “liar’s dividend,” where any genuine recording can be dismissed as fabricated. The semiotic infrastructure that grounded photographic truth—the indexical guarantee—is no longer reliable, and no purely semiotic solution exists. The crisis of indexicality is ultimately a crisis of the codes that governed how we read photographs, and those codes have not yet been replaced.

5.3 · Architectural and spatial semiotics: buildings as sign systems

A building communicates before anyone enters it. Its facade, proportions, materials, and relationship to surrounding space constitute a sign system that viewers decode—often unconsciously—using culturally embedded architectural codes.

Eco’s framework. Umberto Eco, in “Function and Sign: Semiotics of Architecture” (1968), applied his semiotic theory to the built environment by distinguishing primary function (denotation) from secondary function (connotation). A staircase denotes “the possibility of going up”—this is its primary functional meaning, decoded through an architectural code that links the form (ascending horizontal planes) to the use (climbing). But the same staircase may connote grandeur, ceremony, or social hierarchy—these are secondary functions that rest upon but exceed the primary. A throne denotes sitting but overwhelmingly connotes sovereignty; its connotative function is so dominant that it actually impairs its denotative one (thrones are notoriously uncomfortable). Eco’s insight is that architectural meaning is always double: every functional element simultaneously signifies its use and connotes cultural values. A glass facade denotes transparency and light; it connotes corporate modernity and openness. A stone fortress denotes defense; it connotes permanence and power.

Power and space. Institutional architecture systematically deploys spatial semiotics to encode authority. A courthouse elevates the judge’s bench above all other positions, channels movement through restricted zones, and uses heavy materials—marble, stone, dark wood—to connote gravity and permanence. The spatial hierarchy is isomorphic with the legal hierarchy: the physical arrangement is the power relation, made material. A Gothic cathedral uses vertical emphasis—vaulted ceilings, soaring spires, clerestory light—to connote spiritual aspiration and encode human subordination to the divine; the nave channels movement toward the altar as narrative channels attention toward climax. A corporate headquarters lobby uses scale, security checkpoints, and controlled sightlines to signify gatekeeping: access is possible but regulated, and the regulation itself communicates corporate power.

The Panopticon as semiotic architecture. Jeremy Bentham’s Panopticon prison (designed c. 1785) and Michel Foucault’s analysis of it in Discipline and Punish (1975) provide the purest example of architecture as a signifying machine. The design—a central observation tower surrounded by a ring of individually illuminated cells, with shuttered windows that prevent inmates from seeing whether a guard is present—produces a sign of surveillance that operates independently of actual surveillance. The prisoner cannot verify whether they are being watched at any given moment and must therefore assume permanent visibility. The building’s form means “you are observed,” and this meaning produces its intended effect—self-regulation—regardless of whether any observer is actually present. Foucault saw in this the mechanism of modern disciplinary power: control achieved not through spectacular punishment but through the internalization of the gaze. “He who is subjected to a field of visibility, and who knows it, assumes responsibility for the constraints of power; he makes them play spontaneously upon himself.” The architectural sign replaces the sovereign’s sword. The contemporary relevance is obvious: digital platforms replicate panoptic logic when they see users while remaining opaque themselves, producing the same asymmetry of visibility that Bentham engineered in stone.

Space syntax. Bill Hillier’s space syntax, developed at University College London from the late 1970s, provides mathematical tools for analyzing how spatial configuration controls movement and social interaction. The core insight is that spatial structure itself generates movement patterns—you do not need to know what activities occur in a space to predict how people will move through it. Hillier demonstrated that integration (a measure of how topologically close a space is to all other spaces in a network) predicts pedestrian flow: highly integrated streets attract more movement, and movement attracts commercial activity. Space syntax reveals that spatial design is never neutral. A housing estate with a tree-like spatial structure (deep, segregated) produces different social outcomes than one with a grid-like structure (shallow, integrated). The semiotics of public versus private space emerges directly from configurational properties: public spaces tend toward high integration and connectivity; private spaces toward depth and segregation. The boundaries between them are maintained by semiotic markers—gates, thresholds, changes in material, “No Entry” signs—that signal the transition between coding regimes.

5.4 · Biosemiotics: signs beyond the human

The extension of semiotics beyond human communication is not a metaphorical conceit but a theoretical commitment with rigorous conceptual foundations. Biosemiotics argues that semiosis—the production and interpretation of signs—is coextensive with life itself.

Uexküll’s Umwelt. The foundational concept is Jakob von Uexküll’s Umwelt (1934): the subjective, species-specific world of meaning constituted by an organism’s own sign relations. The Umwelt is not the physical environment (Umgebung) as described by an outside observer but the particular perceptual and effectual world that the organism itself constructs through its sensory apparatus and behavioral repertoire. Uexküll’s famous example is the tick, whose Umwelt is constituted by exactly three carriers of significance: the odor of butyric acid (emitted by mammalian skin), the temperature of 37°C (mammalian blood heat), and the tactile quality of mammalian hair. From the billions of features that constitute the tick’s physical surroundings, only these three exist in its Umwelt—everything else is semiotically null. The tick does not inhabit a simplified version of the human world; it inhabits a different world, constituted by different sign relations.

The mechanism connecting organism to environment is the functional circle (Funktionskreis). Receptor organs detect an environmental feature and encode it as a perception sign (Merkzeichen), which triggers effector organs to produce an action sign (Wirkzeichen) that modifies the environment, which in turn alters the available perception signs, completing the circle. This is not linear stimulus-response but a recursive loop of sign production and interpretation—what we would now recognize as a feedback system, though Uexküll described it twenty-five years before cybernetics formalized the concept. Every organism, from tick to human, is connected to its world through functional circles, and the totality of these circles constitutes its Umwelt. Uexküll’s point was Kantian: organisms do not access reality-in-itself but constitute their own phenomenal worlds through species-specific sign relations.

Sebeok’s extension. Thomas Sebeok expanded Uexküll’s framework into a general biosemiotic program, arguing that semiosis is the criterial attribute of life—wherever there is life, there is sign activity. This is not a loose analogy but a precise claim: living systems maintain themselves by producing, transmitting, and interpreting signs, and this semiotic activity is what distinguishes them from non-living physical systems. Sebeok fought against what he called glottocentrism—the restriction of semiotics to human language—insisting that language is a special (and relatively recent) case of a far more ancient biological capacity.

The immune system as sign interpreter. The most striking support for biosemiotics comes from immunology. The immune system’s fundamental operation is discriminating self from non-self—recognizing molecular markers on cell surfaces and classifying them as belonging to the organism or foreign to it. This is an irreducibly semiotic operation: immune cells “read” molecular signs (antigens as representamens in Peircean terms), which stand for potential threats (objects), and produce immune responses (interpretants). The reading is not merely mechanical pattern-matching; the immune system learns, remembers, and adjusts its interpretive criteria over time. Autoimmune diseases represent failures of semiotic discrimination: the system misreads self-markers as foreign, mounting destructive responses against the organism’s own tissues. If we define semiosis as the process by which something functions as a sign for some interpreter, the immune system qualifies without metaphorical stretching.

Evolution as semiotic process. Marcello Barbieri argued that the genetic code is a genuine code, not merely a metaphor: the correspondence between codons and amino acids is arbitrary (maintained by molecular machinery, not by chemical necessity) and conventional (shared across nearly all life). If this is a code in the semiotic sense, then every cell is a semiotic system, and evolution—which operates on the replication and modification of coded information—is a semiotic process. Jesper Hoffmeyer added the concept of code-duality: life is characterized by a fundamental duality between digital codes (genes, which store information) and analog codes (organisms, which act in the world). The interplay between these two coding systems drives evolutionary creativity. Natural selection can then be understood as a kind of reading: the environment “interprets” the phenotypic signs produced by genetic codes and “responds” through differential survival and reproduction. This is not consciousness; it is semiosis at the most basic level—sign-mediated interaction between systems.

What biosemiotics reveals. If semiosis extends to all life, then human meaning-making is not a miraculous exception but the most complex expression of a capacity that pervades the biosphere. This has a humbling implication: the human semiosphere, with its languages, art forms, digital networks, and philosophical traditions, is a spectacular elaboration of processes already present in the tick’s three-sign Umwelt. It also has a challenging one: if Peirce was right that semiosis requires an interpretant, and if biosemiotics is right that cells interpret signs, then interpretation does not require consciousness. The boundaries of mind become far less clear than humanist tradition has assumed.

5.5 · Music semiotics: the hardest case

Music is the most resistant domain for semiotic analysis because it is the least referential of all major art forms. A painting can depict a landscape; a poem can describe an emotion; a film can narrate events. Music, in its “pure” instrumental form, does none of these things. It organizes sound in time without pointing to anything outside itself—or so the tradition of “absolute music” claims. Music semiotics contests this claim, but it must account for why music seems to resist reference.

Nattiez’s tripartite model. Jean-Jacques Nattiez, building on Jean Molino’s work, proposed that musical meaning operates across three distinct levels. The poietic level encompasses the processes of creation—the composer’s strategies, intentions, and context. The neutral (or immanent) level is the physical trace—the score, the recording, the acoustic signal—analyzed in its own structural terms, without reference to producer or receiver. The esthesic level encompasses the processes of reception—how listeners perceive, interpret, and respond to the music. The crucial insight is that these three levels do not necessarily align. A composer may intend one meaning; the structural analysis may reveal different patterns; a listener may hear something else entirely. There is no guaranteed communication chain from producer to receiver. Nattiez replaced the standard sender→message→receiver model with sender→trace←receiver, where poiesis and esthesis converge on the trace from different directions but may never meet. This explains why disagreements about “what music means” are so intractable: participants are often operating at different levels without recognizing it.

Musical topics and cultural codes. If music has meaning, much of it is carried by what Leonard Ratner called topics: culturally conventionalized musical figures that carry specific extramusical associations. A dotted rhythm with triadic fanfares signifies “military” because it derives from actual trumpet and drum patterns; a gentle compound meter with drone basses signifies “pastoral” because it imitates bagpipe music. Kofi Agawu catalogued a “universe” of approximately 27 Classical-era topics including dances (minuet, polonaise, sarabande), styles (brilliant, learned, Sturm und Drang), and character types (amoroso, sensibility). Topics function semiotically as symbols with indexical histories: their current conventional meaning rests on a historical chain of associations linking musical patterns to extramusical contexts. The common perception that minor keys sound “sad” is not an acoustic fact but a cultural code—one so deeply internalized that it feels natural, which is precisely what Barthes would call mythological.

Musical indexicality. Henry Louis Gates Jr.’s concept of Signifyin(g) illuminates how meaning works in African American musical traditions, particularly jazz. Signifyin(g) is “repetition with a signal difference”—taking an existing text (a standard, a riff, a harmonic progression) and transforming it in ways that comment on the original. When John Coltrane performs “My Favorite Things,” the performance cites Rodgers and Hammerstein’s show tune while radically recontextualizing it—the citation is audible, and the difference between citation and original is where meaning resides. This is musical intertextuality as cultural practice, and it demonstrates that music can “mean” through exactly the same mechanisms—citation, transformation, dialogism—that operate in verbal texts.

Does music mean? The honest semiotic answer is that music does not carry propositional content—it cannot assert that “the cat is on the mat” in the way language can. But it does carry culturally coded meaning through topics, indexes, and intertextual relations, and it produces effects (emotional, kinesthetic, social) that are sign-mediated even when they resist verbal paraphrase. Music occupies a space between meaning and affect that exposes the limits of a semiotics built primarily on the model of language. Nattiez proposed that music’s relationship to meaning involves “an infinite and unstable potential for embodying meaning”—a chain of interpretants, in Peirce’s terms, that never terminates in a final, definitive interpretant. This makes music not a failure case for semiotics but a limit case that reveals the openness inherent in all semiosis.

---

6 · Jakobson’s Six Functions of Language

Roman Jakobson’s model of the functions of language, presented in “Linguistics and Poetics” (1960), provides the most systematic framework for analyzing what a communicative act is doing—not what it says but what kind of work it performs. The model is indispensable because it reveals that communication is never just transmission of information; every utterance simultaneously performs multiple functions, and the dominant function determines the character of the communication.

The six factors and their functions. Jakobson begins by identifying six constitutive factors of any verbal communication: an addresser sends a message to an addressee; the message requires a context (referent), a code shared by both parties, and a contact (a physical and psychological channel connecting them). Each factor is the focal point of a corresponding function:

The referential function orients toward context—it conveys information about the world. “Water boils at 100 degrees Celsius” is dominated by the referential function. The emotive function orients toward the addresser—it expresses the speaker’s attitude or emotional state. Interjections (“Bah!”, “Ugh!”) are pure emotive function, but the emotive dimension is present in any utterance that reveals something about the speaker. The conative function orients toward the addressee—it aims to influence the receiver. Imperatives (“Come here!”) and vocatives (“O Romeo!”) foreground the conative function; significantly, as Jakobson noted, imperatives cannot be tested for truth or falsity. The phatic function orients toward the contact itself—it establishes, maintains, or terminates communication. “Hello?” “Can you hear me?” “Uh-huh”—these are not informational but relational, checking that the channel is open. The metalingual function orients toward the code—it uses language to discuss language itself. “What do you mean by ‘semiotics’?” is a metalingual utterance. The poetic function orients toward the message itself—it foregrounds the formal properties of the message for their own sake.

The poetic function. Jakobson’s formulation of the poetic function is among the most important single sentences in twentieth-century linguistics: “The poetic function projects the principle of equivalence from the axis of selection into the axis of combination.” In ordinary language, we select words based on equivalence (choosing among synonyms, near-synonyms, or members of the same paradigmatic set) and combine them based on contiguity (syntactic sequence). In poetic language, the principle of equivalence—similarity of sound, rhythm, grammatical structure—becomes the organizing principle of the sequence itself. Words are combined not only for their syntactic compatibility but for their phonetic similarity (rhyme), their rhythmic equivalence (meter), their grammatical parallelism. “I like Ike”—Jakobson’s own example—derives its force from the phonetic chain /ay/ - /layk/ - /ayk/, where the diphthong /ay/ recurs across all three stressed syllables. The equivalence that normally operates behind the scenes (in the paradigmatic choice of “like” over “love” or “support”) has been projected forward into the syntagmatic chain, becoming the constitutive device of the sequence.

The dominant determines the genre. No utterance is monofunctional; all six functions coexist in every communicative act. But one function typically dominates, and this dominance determines what kind of communication we are dealing with. Scientific prose: referential dominant. Lyric poetry: emotive dominant inflected by the poetic. Advertising: conative dominant. Small talk: phatic dominant. Dictionaries: metalingual dominant. Literary art: poetic dominant. The model’s power is diagnostic: by asking “which function dominates here?” we can characterize any communicative act, from a weather report to a prayer.

Application: “Just Do It.” Nike’s slogan engages all six functions simultaneously. The poetic function dominates: three monosyllables with percussive stress, the rhythmic economy of the phrase making it memorable as form. The conative function is foregrounded through the imperative mood—this is a command directed at the consumer. The emotive function expresses determination and willpower. The referential function gestures vaguely toward athletic achievement. The phatic function sustains the brand-consumer relationship across every encounter. The metalingual function operates through the deliberately vague pronoun “it”—what is “it”? The refusal to specify forces the receiver to negotiate the code, filling “it” with whatever personal meaning makes the slogan work for them. The slogan’s genius is that it maximizes the poetic and conative functions simultaneously, achieving both aesthetic memorability and persuasive force.

---

7 · Intertextuality and the Death of the Author

Every text emerges from and into a world already saturated with other texts. The concept of intertextuality names this condition—not as an occasional literary device (allusion, quotation) but as the fundamental condition of all textual meaning.

Kristeva’s formulation. Julia Kristeva coined the term in 1966, introducing Mikhail Bakhtin’s concept of dialogism to French structuralist thought. Bakhtin had argued that all language is inherently dialogic—every word carries traces of its prior uses, every utterance responds to previous utterances and anticipates future ones. Kristeva radicalized this insight: “Any text is constructed as a mosaic of quotations; any text is the absorption and transformation of another.” This is not a claim about deliberate borrowing but about the condition of signification itself. No text originates from a singular creative consciousness; every text is an intersection of textual surfaces, woven from codes, conventions, discourses, and prior texts that precede any individual author. Kristeva replaced intersubjectivity (meaning produced through dialogue between subjects) with intertextuality (meaning produced through dialogue between texts)—a move that decentered the author as source of meaning.

Barthes’s Death of the Author. Roland Barthes’s 1967 essay drew the implications with characteristic clarity. If every text is a tissue of quotations drawn from innumerable centers of culture, then the Author—conceived as the origin and explanation of the text—is a fiction. “It is language which speaks, not the author.” Barthes replaced the Author with the scriptor, who is born simultaneously with the text and whose only power is “to mix writings, to counter the ones with the others.” The Author limits the text by providing a final signified—a biographical or intentional key that closes interpretation. Removing the Author opens the text to the infinite play of signification. “The birth of the reader must be at the cost of the death of the Author.” This is not an empirical claim that authors do not exist but a methodological claim about how meaning works: significance resides not in origin but in destination, not in the producing subject but in the interpretive encounter.

The writerly and the readerly. In S/Z (1970), Barthes distinguished two modes of textuality. The readerly (lisible) text is consumed passively: it presents a transparent narrative, follows familiar conventions, and positions the reader as a spectator accepting or rejecting a finished product. The writerly (scriptible) text demands active production: it obliges the reader to construct meaning from fragmentary, contradictory, or self-reflexive material, making the reader a co-producer. “Why is the writerly our value? Because the goal of literary work is to make the reader no longer a consumer, but a producer of the text.” The distinction is not binary—Barthes demonstrated, through his 200-page analysis of Balzac’s Sarrasine, that even an apparently readerly text can be read in a writerly way, cracked open through attention to its five codes (hermeneutic, proairetic, semic, symbolic, cultural). The readerly/writerly distinction thus describes not inherent textual properties but modes of reading, which is precisely Barthes’s point: meaning is produced by the reader, not deposited by the author.

Genette’s transtextuality. Gérard Genette, in Palimpsests (1982), systematized the relationships between texts into five categories, providing a more precise taxonomy than Kristeva’s expansive concept. Intertextuality (narrowed to the actual co-presence of one text within another: quotation, plagiarism, allusion). Paratextuality (the relationship between a text and its surrounding apparatus: titles, prefaces, covers, footnotes—elements that mediate between text and reader). Metatextuality (the critical or commentary relationship—one text commenting on another, the “critical relationship par excellence”). Hypertextuality (a later text B grafted upon an earlier text A through transformation or imitation: Joyce’s Ulysses as hypertextual transformation of Homer’s Odyssey; parody, pastiche, adaptation). Architextuality (the most abstract relation—a text’s membership in a genre or type: calling something “a novel” or “an elegy”). Genette stressed that these categories overlap and interact; a single text participates in all five simultaneously. The taxonomy’s practical value is that it gives the analyst specific questions to ask: What is literally quoted here? What framing apparatus shapes reception? What is being commented upon? What earlier text is being transformed? What genre expectations are being invoked or violated?

---

8 · Performativity: When Signs Make Rather Than Mirror

The theories surveyed so far largely treat signs as representations—they stand for something, refer to something, mean something about a world that exists independently of the signification. Performativity theory challenges this representationalist model at its foundations, arguing that signs do not merely describe reality but actively constitute it.

Austin’s speech acts. J.L. Austin, in How to Do Things with Words (1962), began by distinguishing constatives (utterances that describe and can be true or false: “The cat is on the mat”) from performatives (utterances that do something: “I promise,” “I thee wed,” “I name this ship the Queen Elizabeth”). Performatives cannot be true or false, only felicitous or infelicitous—successful or unsuccessful. Austin identified felicity conditions: there must be an accepted conventional procedure, the circumstances and persons must be appropriate, the procedure must be executed correctly and completely, and the participants must have the requisite intentions. When conditions fail, the performative is infelicitous: a marriage performed by someone without authority is a misfire (the act is void); a promise made without intention to keep it is an abuse (the act succeeds formally but is hollow).

Austin then distinguished three dimensions of any speech act. The locutionary act is the act of saying something—producing an utterance with sense and reference. The illocutionary act is what is done in saying it—the force (asserting, promising, commanding, questioning). The perlocutionary act is what is achieved by saying it—the effects on the hearer (persuading, frightening, amusing). Crucially, Austin’s own constative/performative distinction collapsed over the course of his lectures: he came to recognize that all utterances have performative dimensions, and that the constative is simply one type of speech act (the assertive). There is no pure description that is not also, simultaneously, an action.

Butler’s extension: gender as performative. Judith Butler radicalized Austin’s framework in Gender Trouble (1990) and Bodies That Matter (1993). Butler’s central claim is that gender is performative, not expressive. This does not mean that gender is a performance—a role deliberately chosen by a pre-existing subject. Butler is explicit: “The reduction of performativity to performance would be a mistake.” If gender were performance, “one woke in the morning, perused the closet for the gender of choice, donned that gender for the day.” Performativity means something more radical: there is no doer behind the deed. Gender identity does not pre-exist its expressions; it is constituted through what Butler calls “a stylized repetition of acts”—gestures, movements, speech patterns, clothing choices, bodily comportments that, through their incessant repetition within a rigid regulatory framework, “congeal over time to produce the appearance of substance, of a natural sort of being.”

The mechanism is citational: every performance of gender cites and re-cites an established norm. When a doctor announces “It’s a girl!” at birth, this is not a neutral description but a performative—a “forcible citation of a norm” that initiates the process of constituting the child as gendered. The utterance works because it cites an entire history of gendered practices; it derives its power not from the individual speaker’s authority but from the accumulated weight of the convention it invokes. Butler draws on Derrida’s concept of iterability: for any sign to function, it must be repeatable, but every repetition occurs in a new context and therefore introduces difference. This is why subversion is possible. If gender is constituted through repetition, then deviant repetitions—citations that fail, that exaggerate, that displace—can destabilize the norm they cite. Drag, for Butler, reveals that “in imitating gender, drag implicitly reveals the imitative structure of gender itself.” All gender is imitation without an original: “gay is to straight not as copy is to original, but as copy is to copy.”

Semiotic implications. Performativity challenges the entire representationalist framework that underlies most semiotic theory. If signs can constitute the realities they purport to describe—if “It’s a girl” does not report a pre-existing gender but begins to produce one—then the sign/referent relationship is not one of correspondence but of production. This connects directly to Barthes’s analysis of myth: both myth and performativity involve naturalization, the process by which culturally constructed meanings come to appear as natural facts. Myth naturalizes ideology by draining historical specificity from signs; performativity naturalizes gender by concealing the citational labor that produces its apparent substance. In both cases, the semiotic analyst’s task is the same: denaturalization—making visible the mechanism that produces the appearance of the natural.

---

9 · Semiotics as Analytical Practice: The Toolkit

Everything presented so far—the frameworks, the specialized domains, the theoretical debates—converges on a practical question: how do you actually do semiotic analysis? This section provides an answer. It is not a recipe to follow mechanically but a disciplined protocol that, if practiced rigorously, will reveal dimensions of meaning invisible to casual reading.

9.1 · The semiotic attitude

Before applying any specific tool, the analyst must cultivate a particular orientation toward signs—a refusal to take meaning at face value. The semiotic attitude involves four commitments:

Suspend the obvious. The first and hardest move is to resist the naturalness of the sign. When a sign seems to simply “mean what it means,” ask: why this sign and not another? The paradigmatic question—what alternatives were available and rejected—is the single most powerful analytical move in the semiotic toolkit. A news photograph shows a politician behind a podium. Why this angle? Why this moment? Why this framing rather than any of the hundreds of other frames available? Every selection is a meaning-producing exclusion.

Look for what is absent. Signs mean as much by what they exclude as by what they contain. Saussure’s principle of differential value—a sign derives its meaning from its difference from other signs in the system—entails that absence is constitutive of meaning. A recruitment advertisement that shows only young employees signifies through the absence of older workers. A national narrative that celebrates certain founders signifies through the erasure of others. The analyst must develop a trained eye for the structuring absences that shape the meaning of what is present.

Identify the codes. No sign is legible without a code—a system of conventions that connects expression to content. The analyst must ask: which codes make this sign readable, and who has access to them? A medical X-ray is a sign, but it is illegible to someone without the diagnostic code. A subcultural fashion statement is a sign, but it communicates only to those who share the relevant style code. Code identification reveals the social distribution of meaning: some codes are widely shared (the highway code, basic linguistic competence); others are restricted to specific communities (legal jargon, musical notation, internet meme literacy). The distribution of codes is always a distribution of power.

Ask whose interests the sign system serves. Signs are not politically neutral. The question “whose interests does this sign system serve?” connects semiotic analysis to the ideological and material conditions of sign production. Barthes’s mythological analysis, Hall’s encoding/decoding model, and Thompson’s taxonomy of symbolic strategies all insist that sign systems operate within power relations. The analyst must ask: who produced this sign? Under what institutional constraints? For whose benefit? What would have to change for this sign to mean something different?

9.2 · A step-by-step analytical protocol

The following nine steps are not strictly sequential—in practice, analysis moves recursively among them—but they provide a systematic scaffold that ensures no major dimension of meaning is overlooked.

Step 1: Denotative inventory. Begin by cataloguing what is actually there—the material signifiers present in the text. For an image: what objects, persons, settings, colors, spatial arrangements are visible? For a verbal text: what words, grammatical structures, phonetic patterns are present? For a building: what materials, forms, spatial relationships are in play? This step corresponds to Panofsky’s pre-iconographic description and to Barthes’s denotative level. The discipline of exhaustive description prevents the analyst from leaping to interpretation before attending to the material.

Step 2: Paradigmatic analysis. For each significant element identified in Step 1, ask: what alternatives were available? What could have appeared here but didn’t? A paradigmatic analysis maps the system of choices from which the present selection was made. If an advertisement uses blue, the paradigmatic alternatives are every other color—and the choice of blue over red, black, or gold is meaningful precisely because the alternatives exist. This step operationalizes Saussure’s insight that value is differential and relational.

Step 3: Syntagmatic analysis. Examine how the present elements combine and constrain each other. What is the sequential or spatial logic of the arrangement? In a sentence, syntagmatic analysis examines word order and syntactic structure. In an image, it examines spatial composition—what is juxtaposed with what, what narrative sequence is implied, what compositional vectors connect elements. Kress and van Leeuwen’s Given/New, Ideal/Real framework is directly applicable here. Syntagmatic analysis reveals how meaning emerges not from individual signs but from their combination.

Step 4: Code identification. Ask: which codes make this text readable? This includes linguistic codes (what language, what register), visual codes (what genre conventions apply—is this read as a documentary photograph or an art photograph?), cultural codes (what shared knowledge is presumed), and technical codes (what medium-specific conventions are in play). Code identification also asks who has access to these codes: the text’s implied audience is defined by the codes it deploys.

Step 5: Connotative mapping. What second-order meanings attach to the first-order denotative signs? This step applies Barthes’s denotation/connotation model. A leather jacket denotes an article of clothing; it connotes rebellion, toughness, subcultural belonging. A serif typeface denotes letters; it connotes tradition, authority, print culture. Connotations are not subjective associations but culturally structured meaning-chains shared by communities of sign-users.

Step 6: Myth and ideology analysis. What is being naturalized? Whose history is being erased? This step applies Barthes’s concept of myth and the critical tradition of ideological analysis. Look for meanings that present themselves as self-evident, as “just the way things are.” These are the candidates for mythological operation. A “family values” political advertisement that shows only heterosexual nuclear families naturalizes a specific family form as the universal one. The analyst’s task is to reveal the historical, cultural, and political specificity of what appears natural—to show that the sign could have been otherwise.

Step 7: Peircean sign-type analysis. Classify significant elements as iconic (signifying through resemblance), indexical (signifying through causal or existential connection), or symbolic (signifying through convention). This is not merely taxonomic but analytically productive. Iconic elements invite examination of what aspects of the object the sign reproduces and what it omits—every icon is selective, and the selection is meaningful. Indexical elements raise questions about causal traces—is this photograph genuinely indexical (camera present at the scene) or simulated? Symbolic elements raise questions about convention—who established this convention, who maintains it, and what work does it do?

Step 8: Intertextual mapping. What other texts does this text invoke, allude to, transform, or respond to? Apply Genette’s taxonomy: Is there direct quotation or allusion (intertextuality)? How do titles, covers, or framing elements shape reception (paratextuality)? Is any other text being implicitly critiqued or commented upon (metatextuality)? Is this a transformation of an earlier text—a parody, adaptation, or remix (hypertextuality)? What genre expectations does it invoke or violate (architextuality)? Intertextual mapping reveals that no text is an island; meaning is always relational.

Step 9: Audience and decoding positions. Who is the implied reader? What competencies, values, and knowledge does the text presuppose? Apply Hall’s encoding/decoding model: what is the preferred reading (the interpretation intended by producers and structured by the dominant code)? What negotiated readings modify the preferred meaning from specific social positions? What oppositional readings decode the message against its intended meaning? This step reminds the analyst that texts are polysemic—they support multiple readings—but not infinitely so. The structure of the sign system constrains interpretation even as it cannot determine it.

9.3 · Failure modes and common traps

Semiotic analysis can go wrong in characteristic ways, and awareness of these failure modes is itself a form of analytical competence.

Over-reading finds ideological conspiracy in every sign, treating all meaning as instrumental manipulation. Not every red dress in a film is a symbol of passion; not every corporate logo is a weapon of capitalist hegemony. Over-reading produces interpretations that are internally consistent but empirically untethered—unfalsifiable because they admit no counter-evidence. The corrective is to ask: what evidence would make me abandon this reading? If no evidence could, the reading is not analysis but ideology.

Under-reading stops at denotation, accepting the surface meaning as the complete meaning. It mistakes naturalization for truth—it sees the myth but takes it for reality. Under-reading is the default mode of sign consumption: we are all under-readers most of the time, because the social function of codes is precisely to make meaning appear transparent. The corrective is the paradigmatic question: why this sign and not another?

Code imperialism assumes that the analyst’s own cultural codes are universal. A Western semiotician reading color symbolism in a Chinese text through Western color codes will produce nonsense. The corrective is ethnographic humility: always ask whose codes are operative and resist the assumption that your interpretive framework is the only legitimate one.

Ignoring materiality treats signs as if they existed in a frictionless semiotic space, detached from the material conditions of their production, circulation, and reception. A billboard advertisement means differently from the same image on a phone screen, not because the signifiers have changed but because the material conditions of encounter—scale, context, duration of exposure, social setting—shape the interpretive process. The corrective is to always ask: where, how, and under what conditions is this sign encountered?

Conflating signifier with sign. A common error is to treat the material signifier as if it were the sign. But the sign, in Saussure’s account, is the union of signifier and signified; in Peirce’s, it is the triadic relation of representamen, object, and interpretant. The physical mark on the page is not the sign—it becomes a sign only in relation to a meaning system. The corrective is relational thinking: meaning is never in the signifier but always between elements of a system.

9.4 · What semiotics cannot do

Intellectual honesty requires acknowledging the limits of the apparatus.

Semiotics cannot determine which reading is “correct” in any absolute sense. It can identify the codes that produce different readings, map the range of possible interpretations, and evaluate which readings are better supported by the textual evidence—but it cannot adjudicate between competing interpretations with finality, because interpretation is itself a semiotic process, subject to the same indeterminacy it analyzes.

Semiotics cannot escape its own sign systems. Every semiotic analysis is itself a text, produced within codes, subject to ideological pressures, open to further analysis. There is no Archimedean point outside semiosis from which to observe signs objectively. This is not a fatal flaw but a structural feature: semiotics is self-reflexive by nature, and the best semiotic analysis acknowledges its own positionality.

Semiotics cannot fully capture embodied, pre-linguistic meaning. Merleau-Ponty’s phenomenological challenge remains potent: the lived body’s relation to the world precedes and exceeds signification. The feeling of vertigo, the taste of a strawberry, the qualitative character of experience resist complete semiotic capture. Semiotics can analyze how these experiences are represented, communicated, and culturally coded, but it cannot reduce them to sign relations without remainder.

Semiotics cannot replace other modes of analysis. It cannot do the work of economics (tracing material flows of capital), psychology (explaining individual cognitive processes), or history (reconstructing causal sequences of events). It operates at the level of meaning-making, which intersects with but does not subsume these other domains.

But semiotics can make visible what other modes leave invisible. Economic analysis can tell you that a product sells; semiotics can tell you why the advertisement works—which codes it activates, which myths it deploys, which subject positions it constructs. Historical analysis can tell you that a propaganda campaign occurred; semiotics can tell you how the signs functioned—what was naturalized, what was excluded, how the coding structured reception. This is its unique contribution: making the mechanisms of meaning available for critical inspection.

9.5 · The field today: live questions

Semiotics is not a completed edifice but a living discipline confronting genuinely open problems. The most pressing arise from the collision between classical semiotic theory and the conditions of twenty-first-century meaning-making.

How does artificial intelligence challenge the sign/interpretant model? Large language models produce grammatically fluent, contextually appropriate, rhetorically effective text. In Peircean terms, they produce signs that generate interpretants in human readers. But the model itself has no interpretant—no experience of meaning, no consciousness of what its tokens signify. Is this semiosis? Peirce’s framework requires that semiosis involve a mind for which the sign functions—but he also described semiosis at levels of generality that do not obviously require consciousness. If biosemiotics is right that cells engage in semiosis without consciousness, then perhaps AI systems do too. Or perhaps the relevant distinction is not consciousness but embeddedness in a world: the cell’s semiosis is connected to survival in an environment, while the language model’s “semiosis” floats free of any existential stake. This is not a settled question but an active frontier.

What is the semiotic status of data? Classical semiotics analyzes signs—entities that stand for something to someone. But the data points that feed machine learning systems do not “stand for” anything in the way a word or image does. A million tagged photographs are not signs in any traditional sense; they are statistical patterns from which algorithms extract regularities. Yet the outputs of these systems function as signs for human interpreters. There is a gap between the input to the machine (data as statistical substrate) and the output it produces (signs for human consumption), and semiotic theory has not yet developed adequate tools for analyzing this gap.

Can semiotics analyze algorithmic meaning-making? The recommendation algorithm that curates a social media feed performs paradigmatic selection (choosing what to show) and syntagmatic combination (arranging it in sequence)—the two fundamental operations of sign production. But it does so without reference to semantic content, using instead behavioral signals (clicks, dwell time, shares) as proxies for relevance. The algorithm produces meaning effects (filter bubbles, radicalization, aesthetic homogenization) without producing meaning in the intentional sense. Semiotics needs new tools for analyzing these unintended but structurally determined meaning effects—a semiotics of the algorithm that goes beyond treating the algorithm as a black box.

How does digital culture transform the code/message relation? Classical semiotics assumes relatively stable codes shared by communities of sign-users. But digital platforms create conditions of hyperspeed code evolution: meme formats emerge, become universally legible, and become obsolete within weeks. The relation between code and message becomes turbulent—messages circulate faster than the codes needed to decode them can stabilize. This may require a more dynamic theory of code change than classical structuralism, with its emphasis on synchronic systems, can provide.

Biosemiotics and the posthuman. If semiosis extends to all life, then the humanist assumption that meaning is a distinctively human achievement cannot be sustained. This does not diminish human semiosis—language, art, mathematics, and digital networks remain extraordinarily complex elaborations of semiotic capacity—but it repositions them within a continuum that includes cellular communication, immune recognition, and ecological sign exchange. The philosophical implications are significant: if meaning pervades the biosphere, then the opposition between “nature” (meaningless mechanism) and “culture” (meaning-bearing human creation) dissolves. What takes its place is an open question, one that connects semiotics to ecology, evolutionary theory, and the ongoing reassessment of human exceptionalism.

These questions are not decorative add-ons to a completed theory. They are the places where semiotics is being actively remade by the conditions it seeks to analyze. The sign—that deceptively simple concept with which we began—turns out to be an inexhaustible object of inquiry, because it is the medium through which all inquiry, including inquiry into signs, takes place. To think semiotically is to recognize that we are always already inside the systems we study—and that this condition, far from being a limitation, is what makes critical analysis both necessary and possible.