pragm

Pragm is a term used in theoretical discussions of practical reasoning to denote a framework for context-aware decision making. It aims to model how agents select actions by weighing goals, constraints, and contextual information, prioritizing outcomes that are directly useful in a given situation rather than purely optimal in abstract terms.

In a pragm framework, core components typically include a representation of context (situational cues, user intent, environmental state), a goal and utility model, and an action-selection mechanism that applies pragmatic constraints such as resource limits and communicative goals. The approach draws on decision theory, natural language understanding, and bounded rationality, and is often described as modular rather than prescriptive.

Relation to other concepts: Pragm is distinct from pragmatics (the study of how context affects meaning) and from pragmatism (a philosophical tradition) though it shares the emphasis on practical consequences. It is sometimes conflated with the term pragma in programming or with pragmatic programming, but pragm is intended to refer to a broader reasoning framework.

Usage and status: The term has appeared in a limited set of theoretical discussions and in speculative AI literature since the mid-2010s. It remains informal and lacks a canonical definition or standard notation. Some scholars use pragm to describe generic pragmatic reasoning modules within larger systems rather than as a standalone theory. See also pragmatics, pragmatism, and pragma.