Skip to content

Conceptual graph

A graph-based knowledge-representation formalism connecting typed concept nodes and relation nodes with a logical interpretation.

Version
v2 · 2026-09-19 · History
Domain-specific #
3827
Domain group
Applied Sciences & Engineering
Origin domain
Computer Science & Software Engineering
Subdomain
Knowledge Representation → Computer Science & Software Engineering
Aliases
CG

Core Idea

Sowa-style conceptual graphs, generic concept maps and property graphs are not interchangeable, and expressiveness depends on the selected CG fragment and projection rules. Natural-language or domain assertions are normalized into a bipartite graph of concepts and conceptual relations, then graph projection, rules or translation to first-order logic supports inference. The abstraction is therefore identified by a declared carrier, a transformation or constraint over that carrier, and an invariant that tells an analyst whether the named structure is genuinely present.

How would you explain it like I'm…

Bubble-and-Box Sentence Pictures

Imagine drawing 'The cat sits on the mat' as a picture: a bubble for 'cat', a bubble for 'mat', and a little box in between that says 'sits on', with arrows joining them. A conceptual graph is a drawing like that, made with strict rules. Because the rules are strict, a computer can use the drawing to figure out new things.

Fact Maps a Computer Can Read

A conceptual graph is a way of drawing the meaning of a sentence or fact so a computer can reason with it. It uses two kinds of shapes: one kind for things or ideas (like "cat" or "mat"), and another kind for how they are related (like "sits on"). Lines only ever connect a thing to a relation, never a thing straight to a thing. The computer can then match these drawings against each other or follow rules to work out new facts. It is a specific system with strict rules, not just any mind map.

Sowa's Concept-Relation Graphs

A conceptual graph, in the style developed by Sowa, turns natural-language or domain statements into a graph with two kinds of nodes: concept nodes (each with a type and possibly a specific referent, like Cat: Tom) and relation nodes (like On or Agent), with edges only between the two kinds. This is called a bipartite graph. The types come from a declared ontology, relations have a fixed number of arguments, and there is notation for contexts and quantifiers. Inference works by graph projection (matching one graph into another), by rules, or by translating the graph into first-order logic. Conceptual graphs are not the same as informal concept maps or property graphs, and what they can express depends on which fragment and rules are used.

 

A conceptual graph, in Sowa's sense, is a knowledge-representation structure in which natural-language or domain assertions are normalized into a bipartite graph of concept nodes and conceptual-relation nodes. Concept nodes carry a type and a referent; relation nodes have a fixed arity and connect only to concept nodes. The formalism rests on an ontology (type hierarchy) of concept and relation types, notation for contexts and quantification, and a logical semantics, often given by translation to first-order logic. Inference proceeds by graph projection, rule application, or reasoning over the logical translation, and canonical formation rules and constraints govern which graphs are well formed. Expressiveness depends on the selected CG fragment and the projection rules adopted. Sowa-style conceptual graphs, generic concept maps, and property graphs are not interchangeable: concept maps lack formal semantics, and property graphs lack the bipartite concept/relation syntax and CG operations. A graph qualifies as a conceptual graph only if these components are explicit, not merely because it draws concepts and links; comparison with neighbors such as concept maps and RDF clarifies the boundary.

Scope of Application

Conceptual graph belongs to knowledge representation and is useful where the analyst can specify the typed knowledge representation carrier, including objects, relations, parameters, conventions, evidence, boundaries, and comparison targets, then evaluate the ontology of concept and relation types, concept nodes with referents, relation nodes and arity, bipartite graph syntax, context and quantification notation, logical semantics or first-order translation, projection and rule operations, canonical formation and constraints and comparison with concept maps and RDF are explicit.

Clarity

The abstraction clarifies a crowded vocabulary by making the ontology of concept and relation types, concept nodes with referents, relation nodes and arity, bipartite graph syntax, context and quantification notation, logical semantics or first-order translation, projection and rule operations, canonical formation and constraints and comparison with concept maps and RDF are explicit the center of the account. A claim should name the carrier, the governing operation or relation, the applicable assumptions, and the recognition test.

Manages Complexity

Without the abstraction, an analyst must reason directly over many local details: the carrier roles, admissibility assumptions, competing conventions, derived invariants, boundary cases, and proof or validation obligations specific to Conceptual graph. Conceptual graph compresses them into the roles in the structural signature. That compression permits comparison across instances without erasing the variables that determine validity. It also exposes which details may be varied safely and which are constitutive.

Abstract Reasoning

  1. Identify the carrier. State what the elements, states, objects, or observations are: the typed knowledge representation carrier, including objects, relations, parameters, conventions, evidence, boundaries, and comparison targets. Reject examples whose alleged carrier belongs to a different problem.
  2. Lock the constitutive rule. Express the ontology of concept and relation types, concept nodes with referents, relation nodes and arity, bipartite graph syntax, context and quantification notation, logical semantics or first-order translation, projection and rule operations, canonical formation and constraints and comparison with concept maps and RDF are explicit independently of one notation or implementation.

Knowledge Transfer

Knowledge transfers strongly among subfields of knowledge representation because they reuse the typed knowledge representation carrier, including objects, relations, parameters, conventions, evidence, boundaries, and comparison targets, Natural-language or domain assertions are normalized into a bipartite graph of concepts and conceptual relations, then graph projection, rules or translation to first-order logic supports inference., and type the carrier, state every parameter and convention in the definition, test that the ontology of concept and relation types, concept nodes with referents, relation nodes and arity, bipartite graph syntax, context and quantification notation, logical semantics or first-order translation, projection and rule operations, canonical formation and constraints and comparison with concept maps and RDF are explicit, compare the nearest accepted identity, and report counterexamples, uncertainty, and limiting cases.

Relationships to Other Abstractions

Local relationship map for Conceptual graphParents appear above the current abstraction, mutual partners to the right, and children below. Node labels state whether each abstraction is prime or domain-specific; colors identify relation types.Conceptual graphDOMAINPrime abstraction: Representation — is a kind ofRepresentationPRIME

Current abstraction Conceptual graph Domain-specific

Parents (1) — outgoing typed relations

  • Conceptual graph is a kind of Representation Prime

    It remains its own entry because its identity is fixed by the ontology of concept and relation types, concept nodes with referents, relation nodes and arity, bipartite graph syntax, context and quantification notation, logical semantics or first-order translation, projection and rule operations, canonical formation and constraints and comparison with concept maps and RDF.

Hierarchy path (1) — routes to 1 parentless root

Neighborhood in Abstraction Space

Conceptual graph sits in a crowded region of the domain-specific corpus (12th percentile for distinctiveness): several abstractions share nearly its structure, so a description that fits it tends to fit its neighbors too.

Family — Semantic Integration & Knowledge Modeling (13 abstractions)

Nearest neighbors

Computed from structural-signature embeddings · 2026-10-25