Semantics

Lore & Background

Semantics examines meaning in language, limited to the meaning of linguistic expressions. It concerns how signs are interpreted and what information they contain, such as dictionary definitions. Linguistic meaning can be analyzed on different levels: word meaning (lexical semantics), sentence meaning (phrasal semantics), and utterance meaning (pragmatics). Semantics is primarily interested in public meaning, as opposed to speaker meaning. Meaning is often analyzed in terms of sense and reference, also called intension and extension. Sense is given by ideas and concepts associated with an expression, while reference is the object to which an expression points. Semantics contrasts with syntax, which studies rules for grammatically correct sentences, and pragmatics, which investigates how people use language in communication. Together with syntactics and pragmatics, semantics is part of semiotics.

Reader's Guide

Semantics is a foundational field within linguistics, semiotics, and philosophy, providing systematic frameworks for understanding how language conveys meaning. Its branches include lexical semantics (word meaning), phrasal semantics (sentence meaning through compositionality), formal semantics (using logic and mathematics), cognitive semantics (psychological perspective), and others like conceptual, computational, and cultural semantics. Theories of meaning include referential theories, ideational theories, causal theories, truth-conditional semantics, verificationist theories, the use theory, and inferentialist semantics. The study of semantic phenomena began during antiquity but was not recognized as an independent field until the 19th century. Semantics is relevant to formal logic, computer science, and psychology. It has both an internal side (connection between words and mental phenomena) and an external side (how words refer to objects and conditions for truth). The word semantics originated from the Ancient Greek adjective semantikos, meaning 'relating to signs', and entered English from the French term semantique, introduced by linguist Michel Bréal at the end of the 19th century.

Did You Know?

• Semantics contrasts with syntax, which studies rules for grammatically correct sentences, and pragmatics, which investigates how people use language in communication.
• The word semantics originated from the Ancient Greek adjective semantikos, meaning 'relating to signs'.
• Lexical semantics studies word meaning, while phrasal semantics studies the meaning of sentences through compositionality.
• Semantics, together with syntactics and pragmatics, is a part of semiotics.

The Mathematical Heart of Meaning

Denotational semantics, originally termed mathematical semantics or Scott–Strachey semantics, stands as one of three principal formal approaches to pinning down what programming languages actually mean, alongside axiomatic and operational semantics. Its central ambition is to identify mathematical structures—called domains—that faithfully capture what a program or program phrase accomplishes. Rather than tracing step-by-step execution or listing logical assertions, this approach assigns each syntactic construct a mathematical object, a denotation, that encodes its behavior. Programs might be modeled as partial functions mapping inputs to outputs, or even as interactive games played between an environment and a system. A foundational commitment running through the entire framework is compositionality: the denotation of any compound expression must be constructible from the denotations of its constituent parts. This principle ensures that understanding a whole program is not a matter of holistic intuition but a systematic assembly of precisely defined pieces, each contributing its mathematical signature to the larger structure.

Strachey, Scott, and the Birth of a Discipline

The intellectual roots of denotational semantics trace back to the collaborative work of Christopher Strachey and Dana Scott, whose foundational publications appeared in the early 1970s. In their original formulation, the meaning of a computer program was conceived as a straightforward function mapping inputs to outputs—a clean, elegant picture that captured the essence of computation for simple, non-recursive cases. The deeper challenge of giving meaning to recursively defined programs demanded a richer mathematical setting. Scott's key insight was to work with continuous functions operating between domains, specifically complete partial orders, which provided the structural machinery needed to reason about self-referential definitions. From this early foundation, the field has steadily expanded its scope. Researchers have pursued appropriate denotational treatments for increasingly sophisticated language features: sequentiality, concurrency, non-determinism, and local state. The approach has been successfully developed for modern languages such as Concurrent ML, CSP, and Haskell, demonstrating that the compositional framework established decades ago remains robust enough to handle the complexities of contemporary programming.

Building Meaning from Nothing: Recursion and Fixed Points

One of the most technically elegant aspects of denotational semantics is how it assigns meaning to recursive definitions. The meaning of a program phrase is modeled as a function from an environment—holding the current values of free variables—to the phrase's denotation. For a simple expression like n times m, supplying an environment where n is 3 and m is 5 yields the denotation 15. The far more interesting case is recursion, illustrated by the factorial function. Rather than defining factorial all at once, the denotation is constructed as the limit of a sequence of partial approximations. The zeroth approximation is the completely undefined function. The first adds the pair (0,1), since computing factorial of zero requires no recursive call. Each subsequent approximation adds one more ordered pair: F applied five times yields the partial function defined at arguments 0 through 4. These partial functions, ordered by inclusion, form a chain-complete partial order. Because each approximation is contained in the next, the Bourbaki–Witt fixed-point theorem guarantees a limit exists. That limit—the least upper bound of the entire chain—is precisely the total factorial function, defined for every natural number.

Taming Concurrency, Exceptions, and Shared State

As programming languages grew more expressive, denotational semantics faced the formidable task of capturing behaviors that defy simple input-to-output mapping. Modern languages like Concurrent ML, CSP, and Haskell introduce concurrency, exceptions, and shared mutable state, all while preserving the compositional principle that the meaning of a phrase is built from the meanings of its subphrases. Consider an applicative expression f(E1, E2): its denotation is defined in terms of the denotations of f, E1, and E2. But in a concurrent setting, E1 and E2 may be evaluated simultaneously, and the execution of one can influence the other through interactions with shared objects, meaning their denotations become mutually dependent. Complicating matters further, either subexpression might raise an exception that abruptly terminates the other's execution. These interactions—where the meaning of a part is no longer independent of its sibling—represent special cases within the broader denotational framework. The compositional structure still holds, but the mathematical objects required to capture these entangled behaviors are considerably more intricate than the simple partial functions of the early theory.