The purpose of the  4D Composite Knowledge Graph is to layer structured spatio-temporal concepts onto the Reflexive Knowledge Graph,

The  4D Composite Knowledge Graph integrates the capabilities of Reflexive Knowledge Graphs (entities, relationships,   Holonymy Relation and multi-level instance of Entity) with those of Compositionality (composite structures) resulting in the addition of the following key features:

1) Mereological  Aggregate Holonymy relationships, providing dynamic locality to composition.

2) Individual-Property distinction: a clear ontological separation between Bounded Individuals (entities that exist in space and time) and their properties (Property Value). A dedicated Typology characterizes how individuals relate to and instantiate these properties.

3) Temporal Framing: the introduction of temporal connections (Temporal Bounding) that define the temporal boundaries of Bounded Individuals, enabling the explicit modeling of state changes, events, and persistence over time.

The  Abstract Knowledge Graph defines the top level constructs of Entitys and Relations which have no spatiotemporal extension and do not participate in composite structures..

Most "data models" function as  Abstract Knowledge Graphs, offering limited semantic expressivity.
In  SysFEAT,  Abstract Knowledge Graphs are used to define all syntactic constructs required for model management, including  Packaging and Annotation (see Annotation Block).

Full semantic capabilities are provided by 4D Composite Knowledge Graphs.

The purpose of the  Compositionality family of concepts is to provide syntactic mechanisms that support the construction of composite Building Blocks.

The  Compositionality pattern handles situations where Building Blocks have an internal structure. Such Building Blocks are called Aggregate Blocks. They are made of internal nested Entitys called Aggregate Members. Aggregate Members contextualizes Building Blocks living outside the Aggregate Block structure (specialization in OMG - Kernel Modeling Language (KerML) is such an Aggregate Member).

Advanced Aggregate Block features are provided by Bounded Aggregates which hide their internal structure to the outside world by means of a formal boundary (see Ordering Connection).

The  Packaging domain defines the core elements of Building Block and Container which provide the basis for building modules.

 Building Blocks are reusable units They have an independent existence. Thereby, they cannot be nested in other structures which would hide their existence (lexical scope). Because of their independent existence, they must be owned by an independent artifact which sole purpose is model management Containers . Containers are dedicated to model management:

. They own blocks to which they provide existence.

. They can provide namespace to blocks.

. They have dependencies on other Containers which blocks have relationships to blocks they packaged.

The  Predication Substrate defines the compositional system of Element, Class of Element, and local predicates (Linkage) from which  SysFEAT's ontological structures are grown. It provides the internal structure of predication - classification, specialization, powertyping, and Compositionality - as a single, universe-stratified, mechanically verified engine that every higher layer inherits.

Unlike predicate logic (where a predicate is not a term) and graph theory (where an edge is not a node), a Linkage is simultaneously the mechanism of relating an Element that can be classified, specialized, and related to other entities. This self-applicability is what allows the framework to provide a theory of Relation - not just relations, but the ability to classify relations, specialize them, and compose them, using the same machinery that applies to the Entitys they relate.

The  Reflexive Knowledge Graph domain defines the top level constructs of Entitys, Relations, their multi-level classification and mereological relationships.

It forms the second layer of  SysFEAT's graph architecture. In this domain, Relations are refied as first class Elements, allowing them to be further described and qualified.

Multi-level classification is implemented by  PowerTyping, which enables open and flexible meta-modeling capabilities, This includes support for reflexive meta-modeling,  for example allowing Meta-Class to be an instance of itself.

Powertyping also applies to relationships, including the  Holonymy Relation.

This design grounds the  Reflexive Knowledge Graph in non-well-founded set theory, ensuring consistency in circular and self-referential structures (see associated external references).

The 4D Composite Knowledge Graph provides additional structuring of Entitys and Relations by implementing the Compositionality pattern which establishes the concepts of dynamic locality and connections.

The  Upper Ontology domain encompasses five fundamental domains that establish the foundation for the modeling syntax and semantic.

1) The Predication Substrate defines basic constructs for relating (Linkage) and classifying Elements from which ontological structures are grown.

2) The Reflexive Knowledge Graph defines the elementary constructs of Entity, Relation,  their multi-level classification and mereological relationships. Its provides the foundation for open meta-modeling.

3) Packaging, as aspect of modularity, defines syntactic constructs used to group reusable entities (Building Blocks) into modules called Containers.

4) Compositionality, another aspect of modularity, defines the syntactic constructs used to build Entitys that have an internal structure and boundaries.

5) The 4D Composite Knowledge Graph defines Bounded Individuals (entities that exists over space and time) how they are composed (aggregate mereology), qualified (properties) and connected, enabling effective representation of meaning.