Literature Review and Theoretical Positioning

Correspondence Architecture is not proposed as a replacement for existing architecture or organization theories. It is proposed as an integrative extension that makes representation correspondence the architectural object. The relevant literatures each explain part of the problem. Their limitations become visible when they are evaluated against one question: how does this tradition explain the preservation of meaningful relationships among organizational representations over time, especially when AI participates in their production and interpretation?

2.1 Enterprise Architecture

Enterprise Architecture has long sought to improve enterprise coherence by representing business capabilities, information systems, processes, applications, data, technologies, principles, standards, and transformation roadmaps (Lankhorst, 2017; Ross et al., 2006; Zachman, 1987). Its central strength is disciplined representation of complex enterprises. It offers ways to connect strategy to systems, describe target states, govern technology decisions, and coordinate transformation across large organizational landscapes.

For Correspondence Architecture, Enterprise Architecture contributes the idea that explicit architecture matters. Organizations cannot rely on ad hoc local decisions if they want enterprise coherence. However, Enterprise Architecture usually treats organizational meaning indirectly. It can model capabilities, processes, applications, and data flows, but it does not make correspondence among representations of intent, decision, work, evidence, outcome, and learning its central architectural obligation. It can document alignment, but alignment is not the same as correspondence. Alignment often asks whether systems support strategy. Correspondence asks whether the relationships among organizational representations remain meaningful, traceable, and revisable as the organization changes.

2.2 Organizational Design

Organizational Design explains structure, roles, decision rights, authority, coordination, routines, and governance mechanisms (Galbraith, 1974; Lawrence and Lorsch, 1967; Venkatraman, 1989). It is indispensable for understanding how organizations allocate attention, distribute work, and make action possible. It also clarifies why architecture cannot be only technical. Organizations are social systems whose structures shape what is seen, decided, remembered, and ignored.

Yet Organizational Design is not primarily a theory of representation networks. It can explain who has authority to decide, how teams coordinate, and how structures support strategy, but it does not specify how representations of intent, decision, work, evidence, outcome, and learning should remain semantically connected over time. Correspondence Architecture therefore extends Organizational Design by treating representational relationships as part of organizational architecture rather than as secondary documentation.

2.3 Systems Theory, Cybernetics, and Sociotechnical Systems

Systems theory and cybernetics contribute a language of interdependence, feedback, regulation, adaptation, and recursive control (Ashby, 1956; Bertalanffy, 1968). These traditions help explain why organizational action cannot be understood by isolating components. Feedback matters because systems must compare current conditions with desired conditions and adjust behavior. Sociotechnical systems theory adds that organizational performance arises from joint social and technical configuration rather than from technology alone (Orlikowski, 2007; Trist and Bamforth, 1951).

Correspondence Architecture draws on these traditions but shifts the focus from feedback as control to correspondence as preserved meaning. Feedback can inform action without ensuring that the representations being compared are semantically traceable. A metric may trigger adaptation while no longer representing the original intent. A control loop may optimize work while weakening evidence. A sociotechnical configuration may coordinate action while leaving learning disconnected from future decisions. Correspondence Architecture therefore treats feedback as necessary but insufficient. The organization must also preserve the relationships that make feedback interpretable.

2.4 Organizational Information Processing and Coordination Theory

Organizational information processing theory explains how organizations manage uncertainty and equivocality by designing information flows, communication channels, and processing capacity (Daft and Lengel, 1986; Galbraith, 1974). Coordination theory explains how interdependent activities are managed through mechanisms such as scheduling, standardization, mutual adjustment, and shared resources (Malone and Crowston, 1994; Star and Griesemer, 1989). These traditions are important because representation correspondence is partly an information and coordination problem.

However, information processing and coordination theories usually focus on how information is processed and how activities are coordinated, not on how the semantic relationships among representations are governed over time. Correspondence Architecture asks a different question. It does not only ask whether information reaches the right actors or whether activities are coordinated. It asks whether the organization can explain how a decision corresponds to intent, how work corresponds to decision, how evidence corresponds to work, how outcome claims correspond to evidence, and how learning corresponds to outcomes.

2.5 Knowledge Management, Organizational Memory, and Organizational Learning

Knowledge management, organizational memory, and organizational learning are central to the theory (Alavi and Leidner, 2001; Argyris and Schon, 1978; March, 1991; Nonaka, 1994; Walsh and Ungson, 1991). Knowledge management emphasizes creation, storage, transfer, and reuse of knowledge. Organizational memory emphasizes retention of prior experience. Organizational learning emphasizes changes in understanding, routines, and action based on experience.

Correspondence Architecture extends these traditions by treating memory as an architectural capability rather than merely a repository. Storage can preserve artifacts while losing the relationships that make those artifacts meaningful. An organization may store a decision but lose its rationale, preserve a metric but lose its evidentiary limits, archive a lesson but never connect it to future intent. In Correspondence Architecture, organizational memory retains representations, relationships, provenance, lifecycle state, evidence, and learning in forms that remain retrievable, reviewable, and usable for future action.

2.6 Digital Twins, Digital Threads, Ontologies, and Knowledge Graphs

Digital twins, digital threads, ontologies, and knowledge graphs provide powerful ways to represent systems, lifecycle relationships, semantics, and traceability (Cleland-Huang et al., 2012; Gotel and Finkelstein, 1994; Gruber, 1993; Hogan et al., 2021; Kritzinger et al., 2018; Madni et al., 2019; Singh and Willcox, 2018; Tao et al., 2019). They show that complex systems can be represented through structured models and connected data. They are especially relevant for organizations that need to trace artifacts across time and context.

Their limitation is not technological inadequacy. Rather, the limitation is that representational technologies do not by themselves define the organizational meaning that should be preserved. A knowledge graph can connect nodes without knowing whether the relationship expresses intent, evidence, governance, learning, contradiction, or unsupported inference. A digital thread can preserve lifecycle traceability without explaining which organizational correspondences matter for cognition. Correspondence Architecture provides the theoretical architecture that can guide such runtime projections without being reducible to them.

2.7 Organizational Cognition and AI-Native Organizations

Organizational cognition research emphasizes that organizations interpret, remember, learn, and act through distributed cognitive processes (Hutchins, 1995; Walsh and Ungson, 1991; Weick, 1995). AI-native organization research extends this condition by introducing non-human participants capable of generating, transforming, summarizing, classifying, and recommending representations (Amershi et al., 2019; Faraj et al., 2018; Floridi and Cowls, 2019; Jobin et al., 2019; National Institute of Standards and Technology, 2023; Raisch and Krakowski, 2021). AI changes the speed and scale of organizational cognition. It also changes the risk profile.

In AI-native organizations, the organization may not know which representations were human-authored, AI-generated, AI-summarized, externally sourced, or institutionally approved. It may not know whether a recommendation preserves intent, whether a summary preserves evidence limits, or whether a learning artifact reflects an outcome or merely an inference. Correspondence Architecture addresses this problem by treating AI as a bounded participant in organizational cognition. AI may create, transform, summarize, relate, retrieve, or explain representations, but it must not become the authoritative source of organizational meaning.

2.8 The Rival-Theory Boundary

The need for Correspondence Architecture becomes clearest when adjacent traditions are treated as rival explanations. Enterprise Architecture explains enterprise coherence, but not correspondence preservation among all organizational representations. Organizational Design explains structure and coordination, but not representation-network governance. Knowledge management explains knowledge storage and reuse, but not the architectural preservation of relationships among intent, decisions, evidence, outcomes, and learning. Cybernetics explains feedback, but not semantic traceability. Digital twins and knowledge graphs explain representational technologies, but not the organizational theory of what must correspond. AI governance explains constraints on AI use, but not the architecture through which AI-generated representations participate in organizational cognition.

Correspondence Architecture therefore occupies a distinct theoretical position. It does not replace these fields. It integrates them around a different architectural primitive: preserved correspondence among organizational representations.