. We identify a structural limitation in existing information representation frame- works: relations are not treated as first-class objects, but as secondary con- structs dependent on entities. This design prevents systems from internally representing higher-order structures, such as relations between relations, with- out relying on auxiliary mechanisms including reification, schema extensions, or external transformations. We introduce the Dynamic Distributed Memory Model (DDMM), a unified structural framework in which both entities and relations are represented using a single primitive type, called cells. Through a binary composition operator, DDMM enables the recursive construction of complex structures while ensuring that all generated constructs remain within the same representational space. We formally establish that DDMM is structurally closed under composition, allowing relations and higher-order constructs to be internalized and manipu- lated without leaving the system. We further define structural equivalence to guarantee consistency across different composition paths, supporting flexible yet stable construction of hierarchical representations. Empirical evaluation on structured datasets demonstrates that DDMM can represent heterogeneous data, identify recurring structural patterns, construct hierarchical abstractions, and support dynamic reclassification while preserving key structural invariants. By eliminating the distinction between entities and relations, DDMM pro- vides a fundamentally unified representational paradigm that enables systems to construct, interpret, and operate on their own relational structures, offering a foundation for modeling complex, self-contained information systems.
Rushdi Jazem (Fri,) studied this question.