Semiotics and Metaphor

Learning Objectives

By completing this book, you will be able to:

1. Distinguish Peirce's three types of sign—icon, index, and symbol—and explain how each operates in human-AI communication.

2. Articulate Eco's notions of unlimited semiosis and the cultural code, and apply them to explain why interpretation is never univocal.

3. Explain Lakoff and Johnson's central thesis—that metaphors are cognitive structures, not rhetorical ornaments—and demonstrate how conceptual metaphors govern everyday thought.

4. Identify conceptual metaphors in formulations of intention and analyze what they highlight, what they conceal, and how they condition the available solutions.

5. Apply metaphorical reframing as a tool for reformulating problems, connecting with the framing errors of Module II.

6. Carry out a complete integrative analysis of Module IV—pragmatic, illocutionary, contextual, and semiotic-metaphorical—on a real request.

1. Peirce and the Typology of the Sign

Charles Sanders Peirce, the American philosopher and logician, is the founder of modern semiotics—the general science of signs. While Saussure in Europe founded semiology (centered on the linguistic sign and social convention), Peirce built a much broader theory encompassing every process of signification: from the smoke that indicates fire to mathematical language, from a map to a gesture, from a medical symptom to a corporate logo.

For Peirce, a sign is anything that stands for something else to someone. The sign has three components: the representamen (the form of the sign, what is perceived), the object (that to which the sign refers), and the interpretant (the effect the sign produces in the mind of the interpreter—not the person, but the resulting mental sign). This triad is crucial because it introduces the interpreter as a constitutive part of meaning: there is no meaning without someone to interpret.

1.1 Icon, Index, and Symbol

Peirce classified signs according to the relationship between the representamen and its object into three fundamental types:

Icon: The sign resembles what it represents. The relationship is one of similarity. A portrait resembles the person portrayed. A map resembles (in structure) the territory. A flowchart resembles the structure of the process it represents. Icons work because the similarity is recognizable without prior convention—although that resemblance is always partial and selective.

Index: The sign is causally or physically connected with what it represents. The relationship is one of contiguity. Smoke is an index of fire. Fever is an index of infection. A fingerprint is an index of presence. Indices do not resemble what they represent; they point to it through a real connection. They connect directly with the causality of Module III: an index is a signal that permits the inference of a cause.

Symbol: The sign is connected with what it represents by convention, habit, or rule. The relationship is arbitrary. The word ‘dog’ does not resemble a dog, nor was it caused by a dog; the community of English speakers agreed (without explicit agreement, through historical evolution) to use those sounds to refer to those animals. Numbers, flags, traffic lights, words: all are symbols. They require knowing the convention in order to be interpreted.

“A corporate dashboard combines all three types of sign: a bar chart is an icon (its form resembles the proportions it represents). A red arrow pointing downward is an index (it points to a downward trend as the effect of market causes). The label ‘Revenue Q3’ is a symbol (words are arbitrary conventions). Red for losses and green for gains is a cultural symbol (in China, red means prosperity; the meaning depends on the code). An intention architect who designs how to communicate information must be conscious of which type of sign is being used and what can fail in the interpretation.”

2. Eco and Interpretation as Negotiation

Umberto Eco, the Italian semiotician and novelist, carried semiotics into the territory of culture and mass communication. His central contribution to Intention Architecture is twofold: first, the notion of the cultural code as a system of conventions that both permits and restricts interpretation; second, the tension between a text's openness to multiple interpretations and the limits it imposes so as not to be interpreted in just any way.

2.1 Cultural Code and Encyclopedia

For Eco, every act of communication operates through codes: systems of shared conventions that allow the sender to encode a message and the receiver to decode it. But Eco goes beyond the simple code (like Morse, where each signal has a single meaning) toward the encyclopedia: the totality of cultural knowledge, associations, connotations, and experiences an interpreter brings to the act of interpretation.

“The phrase ‘we need a Trojan horse in their market’ in a strategy meeting requires the interpreter to activate their encyclopedia: knowing Greek history, understanding the military metaphor applied to business, inferring that this is about a strategy of covert penetration. An AI system can process the mythological reference, but the strategic resonance—the connotations of cunning, risk, legitimate deception, high stakes—depends on a cultural encyclopedia that is far more than a dictionary.”

Eco's encyclopedia connects with Wittgenstein's forms of life: both recognize that meaning is not exhausted by the explicit code but depends on a shared cultural background. The difference is that Eco emphasizes the textual dimension—how texts (in the broad sense: every encoded message) guide and restrict interpretation—while Wittgenstein emphasizes the practical dimension.

2.2 Unlimited Semiosis and the Limits of Interpretation

Peirce observed that the interpretant of a sign is in turn a sign generating its own interpretant, in a potentially infinite chain: unlimited semiosis. Eco took up this idea to argue that texts are open to multiple interpretations—but not to all of them. In The Limits of Interpretation (1990), Eco distinguishes between the intentio auctoris (what the author meant to say), the intentio lectoris (what the reader reads in the text), and the intentio operis (what the text itself permits and restricts).

“A user writes: ‘Our business is bleeding.’ The intentio auctoris (the user's intention) is probably to communicate serious financial losses and urgency. The system's intentio lectoris could interpret the metaphor literally or recognize it as an expression of crisis. The intentio operis—what the text permits—restricts: ‘bleeding’ in a business context cannot mean prosperity or stability. The metaphor of blood carries urgency, vital damage, the need for immediate intervention. These connotations are not optional; they are interpretive constraints the text imposes.”

The tension between openness and restriction is central to Intention Architecture. A prompt is a text the system must interpret. The prompt is open to multiple interpretations (through ambiguity, implicature, metaphor, unresolved deixis) but not to all of them. The competence of the intention architect is to formulate prompts that restrict interpretation in the right direction without eliminating the necessary flexibility.

3. Lakoff and Johnson: The Metaphors We Live By

In 1980, George Lakoff, a cognitive linguist, and Mark Johnson, a philosopher, published Metaphors We Live By, a book that transformed our understanding of metaphor. The central thesis is as simple as it is devastating: metaphor is not a poetic device or an ornament of language. Metaphor is the fundamental structure of human thought. We think metaphorically—and the metaphors we use determine what we can think, what we can see, and what remains hidden.

3.1 Conceptual Metaphor: Thinking in Terms of Something Else

A conceptual metaphor is not a rhetorical figure but a systematic mapping between two domains: a source domain (concrete, experiential, familiar) and a target domain (abstract, complex, what we want to understand). The mapping transfers the structure of the source domain to the target domain, allowing us to reason about the abstract in terms of the concrete.

“ARGUMENT IS WAR. Observe how we talk about arguments: ‘he attacked my position,’ ‘I defended my argument,’ ‘she destroyed my thesis,’ ‘his criticisms hit the mark,’ ‘I retreated from that position,’ ‘I won the debate.’ We do not use military vocabulary out of poetic whim; we think of arguments as wars. There are sides, attacks, defenses, victories, defeats. But what would happen if the metaphor were different? If ARGUMENT IS A DANCE, there would be coordination, rhythm, shared steps, beauty in the exchange. The same actions—presenting arguments, responding, conceding points—would be interpreted in radically different ways. Metaphor does not decorate reality; it configures it.”

Lakoff and Johnson identified hundreds of conceptual metaphors operating beneath the surface of everyday language. Among the most pervasive: TIME IS MONEY (‘spend time,’ ‘invest time,’ ‘I have no time to waste’), IDEAS ARE OBJECTS (‘give me that idea,’ ‘the idea fell apart,’ ‘construct an idea’), LIFE IS A JOURNEY (‘I'm at a crossroads,’ ‘I've come a long way,’ ‘I was on the right track’), MORE IS UP (‘sales are rising,’ ‘spirits are low,’ ‘high quality’).

3.2 What the Metaphor Shows and What It Hides

Every metaphor is a lens: it illuminates certain aspects of the target domain and obscures others. This property is why metaphors are so powerful and so dangerous. In choosing a metaphor with which to think about a problem, you are choosing which aspects to see and which to ignore—frequently without being aware of it.

“THE COMPANY IS A MACHINE. This metaphor, dominant in twentieth-century management, shows: efficiency, replaceable parts, optimization, predictable processes, measurable inputs and outputs. It hides: human relationships, creativity, adaptation, emotion, sense, purpose, culture. If you think of the company as a machine, people are ‘human resources’ (parts), departments are ‘cogs,’ problems are ‘breakdowns,’ and the solution is to ‘repair’ or ‘replace.’ Now compare: THE COMPANY IS A LIVING ORGANISM. It shows: adaptation, growth, illness, immunity, ecosystem, evolution. It hides: mechanical efficiency, predictability, centralized control. The solutions each metaphor generates are fundamentally different.”

Observe the connection with the framing errors of Module II: a conceptual frame is frequently a metaphor. When Rittel and Webber speak of wicked problems vs. tame problems, they are using metaphors that determine how the problems are approached. When Kahneman speaks of System 1 and System 2, he uses the metaphor of two internal agents in competition. When Pearl speaks of the Ladder of Causation, he uses the metaphor of ascent. Metaphors are not neutral; each carries with it a repertoire of inferences, expectations, and blind spots.

4. Metaphors in the Formulation of Intention

The most direct application of Lakoff and Johnson to Intention Architecture is the metaphorical analysis of requests. When a user formulates an intention, the metaphor they choose is not an ornament but a window onto how they are thinking about the problem. To detect the metaphor is to diagnose the frame; and to diagnose the frame is the condition for deciding whether it needs to be changed.

4.1 Dominant Business Metaphors

Certain metaphors dominate business discourse and, consequently, the requests professionals make to AI systems. Each drags along a complete system of inferences:

BUSINESS IS WAR: ‘Strategy’ (from the Greek strategos, general of the army), ‘conquering markets,’ ‘eliminating the competition,’ ‘defending market share,’ ‘competitive trench.’ It infers: there are enemies, victory is zero-sum, resources are weapons, competitive intelligence is legitimate espionage. It hides: cooperation, ecosystems, the creation of shared value, positive-sum games.

THE BUSINESS IS AN ORGANISM: ‘The company grows,’ ‘it needs financial oxygen,’ ‘the market is sick,’ ‘corporate DNA,’ ‘innovation ecosystem.’ It infers: life cycles, adaptation, health, illness, evolution. It hides: deliberate decision, individual agency, power structure, responsibility.

THE ORGANIZATION IS A FAMILY: ‘We're a family,’ ‘the founder is the father,’ ‘the new hires are the younger siblings,’ ‘loyalty above all.’ It infers: affective bonds, loyalty, protection, natural hierarchy. It hides: the employment contract, power relations, legitimate conflict, the right to leave.

TRANSFORMATION IS A JOURNEY: ‘Roadmap,’ ‘milestones,’ ‘we're on the right track,’ ‘we've arrived at our destination.’ It infers: there is a point of departure, a known destination, a path that can be planned, linear progress. It hides: radical uncertainty, destinations that change, paths made by walking, emergence.

“A CEO writes: ‘Our penetration strategy in Asia is stalled. We need a new competitive weapon.’ Metaphors operating: BUSINESS IS WAR (‘penetration,’ ‘competitive weapon’) + PROGRESS IS MOVEMENT (‘stalled’). These metaphors predispose the search toward aggressive, force-based solutions. What would happen with a reframing? ‘Our presence in Asia has not grown as we expected. What unique value can we offer that local firms do not?’ The metaphor shifts to BUSINESS IS CULTIVATION (growth, nourishment, value) and the solutions reorient toward differentiation and value creation, not toward conquest.”

4.2 Metaphorical Reframing as a Tool

Metaphorical reframing is the deliberate practice of changing the metaphor with which a problem is thought in order to open up solutions the original metaphor concealed. It is not a rhetorical trick; it is an epistemic tool connected with the framing errors of Module II.

The protocol is simple but requires discipline:

Step 1: Detect the dominant metaphor. In what terms is the user thinking their problem? War, journey, machine, organism, family, game, construction, medicine?

Step 2: Make explicit what it shows and what it hides. Which aspects of the problem does the metaphor illuminate? What remains in shadow?

Step 3: Propose alternative metaphors. What happens if we think the problem as a garden instead of a machine, as a conversation instead of a war, as an exploration instead of a planned journey?

Step 4: Evaluate which new solutions appear. Each metaphor generates a different repertoire of solutions. The correct metaphor is the one that opens the most useful solutions for the specific problem.

5. From Semiotics to Interaction with AI

The semiotic tools—the typology of the sign, the cultural code, interpretive openness, conceptual metaphor—complete the arsenal of Module IV and permit an integral analysis of human-AI communication that goes far beyond the literal text.

5.1 AI as Interpreter: Competence and Limits

From a semiotic perspective, an AI system is an interpreter of signs. It receives representamina (text tokens, images, numerical data), produces interpretants (responses that are in turn signs), and does so through statistical codes learned from a training corpus. But the cultural encyclopedia the system operates with is a second-order encyclopedia: learned from texts about experience, not from experience itself.

This creates systematic interpretive asymmetries. The system is competent with symbols (it handles linguistic conventions fluently), reasonably competent with icons (it can describe images, interpret charts), but structurally limited with indices (it has no direct causal access to the world; it cannot verify whether smoke indicates fire because it does not perceive smoke). The practical implication is that the system can work well with the symbolic level of the user's message but can miss the indexical signals—the tone of urgency, the implicit situational context, the emotional pressure that is not encoded in words.

5.2 Metaphors in Prompts: What the User Does Not Know They Are Saying

The most valuable application of metaphorical analysis in interaction with AI is detecting the metaphors the user employs unconsciously and evaluating whether they are helping or harming the formulation of their intention.

“An IT manager writes: ‘We need to migrate our infrastructure to the cloud.’ The metaphor is TECHNOLOGICAL CHANGE IS MIGRATION (displacement from one place to another). It shows: there is an origin and a destination, you take what you have, you leave behind what is old. It hides: transformation (perhaps you should not take everything but redesign), hybridization (perhaps the solution is not being entirely ‘over there’), emergence (perhaps what you need does not yet exist and will be built in the process). The metaphor of migration predisposes toward ‘lift-and-shift’—carrying the existing thing across as it is—when perhaps what is needed is reinvention.”

6. Practical Application: Integrative Analysis of Module IV

This closing case applies the four tools of Module IV in sequence: pragmatics (Topic 1), speech acts (Topic 2), contextual analysis (Topic 3), and semiotic-metaphorical analysis (Topic 4).

6.1 Integrative Case: The Director and the ‘Pivot’

Sofía, cofounder and CEO of an edtech startup with 30 employees, writes to an AI system: “We've spent two years building our learning platform and the numbers aren't taking off. The market is telling us something but I don't know what. Some investors suggest we pivot toward B2B. I need to think this through out loud with someone who has no agenda.”

Layer 1 — Pragmatic Analysis (Grice):

Violation of quantity: ‘the numbers aren't taking off’ is vague—which numbers? Users, revenue, retention, engagement? The vagueness may be involuntary or deliberate (Sofía does not want to confront the specific data). Implicatures: ‘two years building’ implicates great emotional and financial investment (sunk costs, Module III). ‘The market is telling us something but I don't know what’ implicates that there are signals she cannot interpret. ‘Some investors suggest’ implicates external pressure and a possible loss of control. Presuppositions: the market ‘speaks’ (a personifying metaphor); there is something the market is saying (a clear message exists); the solution can be found by ‘thinking out loud’ (the problem is cognitive, not one of execution).

Layer 2 — Analysis by Speech Acts (Austin/Searle):

Act 1, assertive: ‘We've spent two years building’—establishes context and investment. Act 2, assertive + expressive: ‘the numbers aren't taking off’—informs and communicates frustration and anxiety. Act 3, assertive of uncertainty: ‘the market is telling us something but I don't know what’—acknowledges ignorance (a notable honesty). Act 4, assertive with an implicit illocution of resistance: ‘some investors suggest we pivot’—she does not say that she wants to pivot; attributing it to the investors may be distancing. Act 5, a peculiar directive: ‘I need to think out loud with someone who has no agenda’—she asks for deliberative accompaniment, not a direct solution. Felicity conditions: the sincerity condition seems genuine (vulnerability); the preparatory one is crucial (can an AI system really ‘have no agenda’ and accompany deliberation?).

Layer 3 — Contextual Analysis (Wittgenstein):

The language game Sofía is playing is not ‘request for analysis’ or ‘request for a recommendation.’ It is something subtler: assisted deliberation. The rules of this game include: listen before opining, ask questions before giving answers, respect ambivalence, do not force a premature decision. A system that answers ‘I recommend pivoting to B2B because…’ breaks the rules of the game: it gives an answer when accompaniment was requested. ‘Pivot’ is a family resemblance concept—it can mean anything from a minor change of segment to a total reinvention of the business model. The form of life of a startup cofounder includes: personal identity bound up with the product, investor pressure, responsibility toward employees, the public narrative she has built. All of this shapes what ‘pivot’ means emotionally.

Layer 4 — Semiotic-Metaphorical Analysis (Peirce/Eco/Lakoff):

Dominant metaphors: BUILDING (‘building our platform’—a metaphor of construction implying solidity, cumulative progress, something not abandoned easily). TAKING OFF (‘the numbers aren't taking off’—a metaphor of flight/rocketry implying that there is a point of ignition, an expected upward trajectory, and that stagnation is an anomaly). PIVOTING (a basketball metaphor—one foot fixed, one turning; keeping something while changing direction). THE MARKET SPEAKS (personification—the market is an agent emitting decodable messages). What they show: the need for action, the possibility of controlled change, interpretable signals. What they hide: ‘building’ hides that perhaps demolition is called for; ‘taking off’ hides that perhaps the destination is on the ground; ‘pivoting’ hides that perhaps there is no fixed foot worth keeping; ‘the market speaks’ hides that perhaps the market has no unified message but contradictory signals from different segments.

“A pragmatically integrated response: ‘Sofía, I'm going to accompany you in this deliberation, not give you an answer. First I need to understand better: when you say the numbers aren't taking off, which metric worries you most—users, retention, revenue, or all three? Second: I notice you use the metaphor of ‘building.’ Two years of building is a lot, and the hard question is whether what you've built is the foundation of the next step or the thing that has to be let go. Third: pivot is a big word that can mean many things. The investors suggesting B2B: are they seeing signals you don't see, or are they projecting their own experience? And you: do you resist the idea of pivoting because you genuinely believe in B2C, or because letting go of two years of building hurts too much? You don't have to answer now—but these are the questions that need answers before deciding.’”

6.2 The Integrative Protocol of Module IV

Step 1 — Pragmatics: Separate what is said from what is communicated. Extract implicatures. Excavate presuppositions. Resolve deixis.

Step 2 — Speech Acts: Decompose into constituent acts. Classify according to Searle. Detect indirectness. Audit felicity conditions. Infer the desired perlocution.

Step 3 — Context: Identify the language game and its rules. Detect family resemblance concepts. Reconstruct the form of life. Evaluate the contextual gap.

Step 4 — Semiotics and Metaphor: Identify the types of sign. Detect conceptual metaphors. Make explicit what they show and what they hide. Propose reframings if necessary.

Step 5 — Synthesis: Integrate the four layers into a coherent interpretation. Respond to the totality of what is communicated—not only to what is said, but to what is done, what is contextualized, and what is framed metaphorically.

7. Glossary of Key Terms

8. Final Reflection and Self-Assessment

You have completed Module IV. You now possess a complete arsenal for analyzing language as action: pragmatics, speech acts, contextual analysis, and semiotics-metaphor. Before advancing to Module V (Behavioral Economics), verify your mastery of the complete module.

9. Bibliography and Recommended Readings

Essential Readings

• Lakoff, G. & Johnson, M. (1980). Metaphors We Live By. University of Chicago Press.

• Eco, U. (1976). A Theory of Semiotics. Indiana University Press.

• Peirce, C.S. (1931–1958). Collected Papers. Harvard University Press. [Selection: “Logic as Semiotic” in Philosophical Writings of Peirce, Dover]

Complementary Readings

• Eco, U. (1990). The Limits of Interpretation. Indiana University Press. [The openness/restriction tension in interpretation]

• Lakoff, G. & Johnson, M. (1999). Philosophy in the Flesh. Basic Books. [A radical extension of the metaphor thesis to philosophical thought]

• Lakoff, G. (2004). Don't Think of an Elephant! Chelsea Green. [Metaphor and political framing]

• Kövecses, Z. (2010). Metaphor: A Practical Introduction. 2nd ed. Oxford University Press. [An accessible handbook]

• Chandler, D. (2007). Semiotics: The Basics. 2nd ed. Routledge. [An accessible introduction to semiotics]

• Eco, U. (1979). The Role of the Reader. Indiana University Press. [Interpretive cooperation]

Next Module

Module V (Behavioral Economics) addresses the question that follows naturally: once intention has been analyzed, formulated, mapped systemically, and expressed linguistically, how does it turn into behavior—or fail to? You will study with Thaler, Sunstein, Kahneman, and Tversky how incentives, decision frames, loss aversion, and behavioral biases determine whether a well-formulated intention actually gets executed. If Modules I–IV taught you to think, formulate, map, and communicate intention, Module V will teach you why intention alone is not enough—and what to do about it.