The Architecture of Navigation
GPS helps us navigate places.
What if AI could navigate meaning?
GPS gave machines a reliable way to locate themselves in the physical world. tokum.ai applies the exact same idea to information: give concepts identifiable addresses, map how they connect, and preserve the context needed to interpret them.
The Analogy: Google Maps makes physical locations addressable and traversable. Meaning Navigation makes Identity, relational Maps, and observer Context addressable and traceable across networks.
The Challenge
Fluent answers are not always grounded answers.
Today's AI can produce convincing language, but it can be difficult to tell which statements are verified, which are inferred, and which remain unknown. Better navigation starts with a clearer map of what is known.
Identify
What concept are we talking about? Different words, files, or formats may refer to the same idea. Without stable identities, machines confuse surface phrasing with meaning.
Connect
How does that concept relate to other concepts, evidence, sources, and situations? Like streets linking towns, meaning requires a structured roadbed of relationships.
Respect the Boundary
What can be supported by available evidence—and what is not yet established? When a road ends, a navigator halts; a knowledge system must acknowledge what remains unobserved.
From Physical Coordinates to Semantic Coordinates
Navigation needs more than a destination.
A useful navigation system needs a way to identify where things are, describe how they connect, and interpret the journey from a particular point of view. The same principles help organize knowledge:
Give things an identity
A GPS coordinate identifies a place. A canonical identifier gives a concept a stable reference point that systems can resolve without ambiguity.
Describe relationships
A road map shows physical routes between places. A semantic map describes explicit connections between concepts and the evidence supporting those connections.
Understand the situation
Your route depends on where you are and where you need to go. Interpretation also depends on the observer's specific context, perspective, and purpose.
The Building Blocks
How tokum.ai approaches meaning
The goal is to make meaning easier for people and machines to identify, connect, inspect, and reuse—without losing the context that makes information useful.
Stable Identity
A Canonical Concept Identifier (CCI) provides a consistent address for a concept, rather than relying only on the surface words used to express it.
Context Matters
A Contextual Tokum Instance (CTI) captures a concept as it actually appears in a particular situation, binding the abstract sign to its specific observer and perspective.
Evidence & Limits
Clear provenance helps people inspect where an assertion came from. When a verified answer is not available, the boundary of knowledge (Epistemic Zero) is reported directly.
A Simple Example
One idea. Many ways to express it.
Imagine a knowledge system handling the concept of an apple tree. A person describes it in everyday conversation, a scientist uses a botanical term (Malus domestica), and an inventory database stores an orchard SKU.
The words and formats differ, but the system connects them to a shared concept—while preserving the source, evidence, and context of each statement.
Illustrative Tokum Example
Apple Tree
Identity: A tree that produces apples.
Relationships: Plant, tree, fruit, orchard block.
Context: Botanical taxonomy, farm ledger, or drone sensor reading.
Evidence: The verifiable source supporting each observation.
Unknown (Epistemic Zero): Details that cannot be confirmed from available evidence are left unstated—not invented.
A Map of Relationships
Semantic Spacetime
A map of meaning needs dimensions. In the Semantic Spacetime model used by tokum.ai, four fundamental dimensions describe how concepts are situated in relation to one another:
Proximity
Similarity & Clustering
How closely two concepts or entities are related across conceptual distance.
Sequence
Time & Causality
In what temporal or causal order events, actions, or procedures occur.
Containment
Part-Whole Hierarchy
What belongs within something else, or forms part of a larger whole.
Property
Characteristics & Rules
Which characteristics describe an entity or distinguish it from others.
From Individual Knowledge to Shared Understanding
A connected architecture for knowledge
Knowledge exists at different scales: in a person's private notes, inside a team, and across a wider community. These contexts remain distinct while still allowing permissioned, trusted connections:
Private Perspective
Private Perspective Portal (PPP)
A personal enclave for notes, private records, and individual perspective. Runs securely on edge hardware.
Shared Understanding
Shared Semantic System (SSS)
A bounded context where a team or enterprise can coordinate and connect knowledge with clear consent.
Broader Knowledge
Comprehensive Comprehension Cloud (CCC)
A shared, public directory for connecting canonical concepts, scientific constants, and common references across contexts.
These describe the organizational architecture: private data stays private, and connections are governed by explicit permission.
For the Technically Curious
Meaning with structure and traceability
Behind the navigation analogy is a structured approach to representing concepts, relationships, context, and provenance:
Canonical Identity
Canonical identifiers make concepts directly referable across applications, languages, and data sources, reducing dependence on any single wording or surface format.
Explicit Relationships
Instead of leaving connections implicit in blocks of text, the architecture represents links as structured relationships that can be examined, traced, and verified.
Provenance
Source and context information help users inspect how an assertion entered the knowledge system and what evidence supports it.
Boundaries of Knowledge
Epistemic Zero represents the state in which the system has no verified claim available within the relevant boundary. It formalizes uncertainty rather than guessing.
Explore Core Terminology
Canonical Concept Identifier (CCI) — a canonical mathematical reference coordinate for a concept.
Contextual Tokum Instance (CTI) — a concept represented as observed within a particular situation and context.
Semantic Spacetime (SST) — a relational framework organized around proximity, sequence, containment, and property.
Epistemic Zero (Ø) — the formal representation of an unobserved coordinate or boundary of verified knowledge.
A Different Way to Organize Knowledge
Move from finding words to navigating meaning.
Explore how tokum.ai connects identity, relationships, and context into an open architecture for machine-readable meaning.
The Architecture of Navigation
GPS helps us navigate places.
What if AI could navigate meaning?
GPS gave machines a reliable way to locate themselves in the physical world. tokum.ai applies the exact same idea to information: give concepts identifiable addresses, map how they connect, and preserve the context needed to interpret them.
The Analogy: Google Maps makes physical locations addressable and traversable. Meaning Navigation makes Identity, relational Maps, and observer Context addressable and traceable across networks.
The Challenge
Fluent answers are not always grounded answers.
Today's AI can produce convincing language, but it can be difficult to tell which statements are verified, which are inferred, and which remain unknown. Better navigation starts with a clearer map of what is known.
Identify
What concept are we talking about? Different words, files, or formats may refer to the same idea. Without stable identities, machines confuse surface phrasing with meaning.
Connect
How does that concept relate to other concepts, evidence, sources, and situations? Like streets linking towns, meaning requires a structured roadbed of relationships.
Respect the Boundary
What can be supported by available evidence—and what is not yet established? When a road ends, a navigator halts; a knowledge system must acknowledge what remains unobserved.
From Physical Coordinates to Semantic Coordinates
Navigation needs more than a destination.
A useful navigation system needs a way to identify where things are, describe how they connect, and interpret the journey from a particular point of view. The same principles help organize knowledge:
Give things an identity
A GPS coordinate identifies a place. A canonical identifier gives a concept a stable reference point that systems can resolve without ambiguity.
Describe relationships
A road map shows physical routes between places. A semantic map describes explicit connections between concepts and the evidence supporting those connections.
Understand the situation
Your route depends on where you are and where you need to go. Interpretation also depends on the observer's specific context, perspective, and purpose.
The Building Blocks
How tokum.ai approaches meaning
The goal is to make meaning easier for people and machines to identify, connect, inspect, and reuse—without losing the context that makes information useful.
Stable Identity
A Canonical Concept Identifier (CCI) provides a consistent address for a concept, rather than relying only on the surface words used to express it.
Context Matters
A Contextual Tokum Instance (CTI) captures a concept as it actually appears in a particular situation, binding the abstract sign to its specific observer and perspective.
Evidence & Limits
Clear provenance helps people inspect where an assertion came from. When a verified answer is not available, the boundary of knowledge (Epistemic Zero) is reported directly.
A Simple Example
One idea. Many ways to express it.
Imagine a knowledge system handling the concept of an apple tree. A person describes it in everyday conversation, a scientist uses a botanical term (Malus domestica), and an inventory database stores an orchard SKU.
The words and formats differ, but the system connects them to a shared concept—while preserving the source, evidence, and context of each statement.
Illustrative Tokum Example
Apple Tree
Identity: A tree that produces apples.
Relationships: Plant, tree, fruit, orchard block.
Context: Botanical taxonomy, farm ledger, or drone sensor reading.
Evidence: The verifiable source supporting each observation.
Unknown (Epistemic Zero): Details that cannot be confirmed from available evidence are left unstated—not invented.
A Map of Relationships
Semantic Spacetime
A map of meaning needs dimensions. In the Semantic Spacetime model used by tokum.ai, four fundamental dimensions describe how concepts are situated in relation to one another:
Proximity
Similarity & Clustering
How closely two concepts or entities are related across conceptual distance.
Sequence
Time & Causality
In what temporal or causal order events, actions, or procedures occur.
Containment
Part-Whole Hierarchy
What belongs within something else, or forms part of a larger whole.
Property
Characteristics & Rules
Which characteristics describe an entity or distinguish it from others.
From Individual Knowledge to Shared Understanding
A connected architecture for knowledge
Knowledge exists at different scales: in a person's private notes, inside a team, and across a wider community. These contexts remain distinct while still allowing permissioned, trusted connections:
Private Perspective
Private Perspective Portal (PPP)
A personal enclave for notes, private records, and individual perspective. Runs securely on edge hardware.
Shared Understanding
Shared Semantic System (SSS)
A bounded context where a team or enterprise can coordinate and connect knowledge with clear consent.
Broader Knowledge
Comprehensive Comprehension Cloud (CCC)
A shared, public directory for connecting canonical concepts, scientific constants, and common references across contexts.
These describe the organizational architecture: private data stays private, and connections are governed by explicit permission.
For the Technically Curious
Meaning with structure and traceability
Behind the navigation analogy is a structured approach to representing concepts, relationships, context, and provenance:
Canonical Identity
Canonical identifiers make concepts directly referable across applications, languages, and data sources, reducing dependence on any single wording or surface format.
Explicit Relationships
Instead of leaving connections implicit in blocks of text, the architecture represents links as structured relationships that can be examined, traced, and verified.
Provenance
Source and context information help users inspect how an assertion entered the knowledge system and what evidence supports it.
Boundaries of Knowledge
Epistemic Zero represents the state in which the system has no verified claim available within the relevant boundary. It formalizes uncertainty rather than guessing.
Explore Core Terminology
Canonical Concept Identifier (CCI) — a canonical mathematical reference coordinate for a concept.
Contextual Tokum Instance (CTI) — a concept represented as observed within a particular situation and context.
Semantic Spacetime (SST) — a relational framework organized around proximity, sequence, containment, and property.
Epistemic Zero (Ø) — the formal representation of an unobserved coordinate or boundary of verified knowledge.
A Different Way to Organize Knowledge
Move from finding words to navigating meaning.
Explore how tokum.ai connects identity, relationships, and context into an open architecture for machine-readable meaning.