The informing act: apprehension in AI-mediated systems

Purpose This article formalizes a semantic complement to Shannon’s syntactic theory of communication. Where Shannon bracketed meaning to formalize transmission, this framework brackets transmission to formalize apprehension: the conditions under which a knowing subject comes to know. It provides an epistemic criterion (K ≠ Kʹ) specifying when encounter with difference produces configuration change in a knowing subject. Design/methodology/approach The framework is conceptual, not empirical. It draws on Polanyi’s tacit knowledge to ground the knowing subject, inverts Brookes’ fundamental equation to define the informing act by its criterion rather than assuming information as given and unifies four adjacent threads in information science (Buckland’s document theory, Bates’s situated interpretation, Taylor’s value-added processes and O’Connor and Bonnici’s proximity) under a single formalization. A demonstration drawn from pianist Hélène Grimaud’s memoir instantiates each component of the framework. Findings Three contributions follow: First, a formal definition of the informing act grounded in cognitive effect: apprehension producing K ≠ Kʹ; second, a formalization of information systems as infrastructure for apprehension, where processes add value by engineering proximity rather than producing knowing; third, a triadic interface model in which mediating systems, including Artificial Intelligence (AI), contribute to conditions for apprehension without participating in the informing act itself. Research limitations/implications The framework specifies conditions for the informing act without providing a complete theory of knowledge structure. Granularity of configuration change is domain-dependent and open to empirical investigation. The knowing subject is individual, grounded in tacit integration, deliberately excluding collectives and current AI systems as knowing agents. Practical implications The framework reorients information system design from retrieval metrics toward apprehension outcomes, providing practitioners a criterion for distinguishing when systems serve knowing from when they merely increase output volume. Social implications The framework distinguishes AI-generated output from knowing by specifying that the informing act requires apprehension by a knowing subject. It also clarifies why “information overload” is imprecise: what saturates is input. Originality/value Where Brookes begins with ΔI and describes its effect on a knowledge structure, this framework inverts the relationship: the criterion defines the informing act, not the input. This inversion is not incremental extension but conceptual reorientation, consolidating what the semantic thread in information science has long pointed toward and completing the dyad Shannon’s syntactic formalism left open.