docs.universalhex.org

Universal Hex
Taxonomy

The Universal Hex Taxonomy (UHT) introduces a 32-bit semantic encoding framework that captures the essential traits of any entity — physical, functional, abstract, or social — in a fixed-length binary code. Each bit corresponds to one of 32 clearly defined properties, producing a compact semantic fingerprint. UHT enables precise comparisons via Hamming distance, supports clustering, and remains accessible to both humans and machines.

01 Introduction

No existing system combines universal scope, compactness, and interpretability. Domain-specific taxonomies lack flexibility, ontologies become unwieldy, and neural embeddings offer little transparency. UHT fills this gap with a simple fixed-length code built from binary traits organized into four layers: Physical, Functional, Abstract, and Social.

Every object, concept, or system can be tagged, compared, and understood using the same format. The result is a compact code that captures important characteristics while remaining explainable.

02 Structure & Layers

Each UHT code consists of 32 bits divided into four groups of eight. These correspond to four semantic layers:

Bit Layout

Bit RangeLayerDescription
1–8PhysicalTangibility, materiality, form
9–16FunctionalPurpose, behavior, operation
17–24AbstractSymbolism, logic, structure
25–32SocialCulture, regulation, identity

Each number maps to a trait. Group the bits into four bytes and convert to hexadecimal for readability. The result is an 8-character hex string like CE880000.

Converting Bits to Hex

Example: encoding a Paperclip.

# 1. Write the 32-bit string in groups of eight 11001110 10001000 00000000 00000000 # 2. Convert each byte to hex 11001110CE 1000100088 0000000000 0000000000 # 3. Concatenate UHT Code: CE880000

To decode, reverse these steps: split the hex into byte pairs, convert to binary, and match each bit to its corresponding trait.

03 Trait Reference

Each of the 32 bits corresponds to a named property. Bits are either on (1) or off (0). Below is the complete reference with edge cases from the classification specification.

Physical Layer · Bits 1–8

BitTraitDescription
1Physical ObjectA discrete, bounded physical entity that occupies space and has tangible form.
2SyntheticCreated, manufactured, or significantly shaped by human activity or intention.
3Biological/BiomimeticHas biological origin or structure inspired by biology.
4PoweredRequires external energy input to operate or perform its primary function.
5StructuralServes a load-bearing, shape-maintaining, or structural integrity function.
6ObservableCan be directly detected by human senses or scientific instruments.
7Physical MediumComposed of physical matter — has substance and mass.
8ActiveExhibits autonomous behavior or initiates actions independently.
Edge cases — Physical Layer
1A cloud — While diffuse, a bounded collection of physical water droplets
0Fire/flame — Fire is a process (combustion), not a bounded physical object
1A river — Has physical boundaries and tangible water
0Electricity — A phenomenon/flow of charge, not a discrete object
1Cultured pearls — Human intervention controls pearl formation
0Wild managed forest — Management doesn’t make it synthetic
1Solar-powered calculator — Requires external energy (light) to operate
0Mechanical watch — Powered by manual winding/spring, not external energy
1Computer virus — Executes and propagates autonomously once activated
0Biological virus — Cannot act autonomously, requires host cell machinery

Functional Layer · Bits 9–16

BitTraitDescription
9Intentionally DesignedCreated or adapted with a deliberate purpose or intended function.
10Outputs EffectActively produces measurable outputs: light, sound, force, heat, or information.
11Processes Signals/LogicInterprets, transforms, routes, or computes data, signals, or logical operations.
12State-TransformingCan change its own internal state or configuration based on inputs or time.
13Human-InteractiveDesigned to be directly used, operated, or engaged with by humans.
14System-IntegratedOperates as part of a larger system or network; function depends on context.
15Functionally AutonomousOperates independently without requiring continuous external control.
16System-EssentialCritical to a system’s operation — removal would cause failure or degradation.
Edge cases — Functional Layer
0Spider web — Created by instinct, not intentional design
1Bonsai tree — Intentionally shaped by human cultivator
1Battery — Outputs electrical energy
0Mirror — Redirects light, doesn’t generate it
1DNA — Encodes and processes genetic information
0Enzyme — Catalyzes reactions but doesn’t process information
1Thermostat — Changes between on/off states based on temperature
0Server computer — Not intended for direct human interaction (accessed remotely)
1Elevator — Operates autonomously based on button presses and internal logic
0Appendix (organ) — Can be removed without system failure

Abstract Layer · Bits 17–24

BitTraitDescription
17SymbolicRepresents concepts through signs, models, or conventional meaning.
18SignallingActively transmits or encodes information through signals or media.
19Rule-GovernedBehavior or structure defined by formal rules, algorithms, or logical systems.
20CompositionalStructured in meaningful layers, modules, or hierarchical components.
21NormativeDirects, guides, or constrains behavior through rules or expectations.
22MetaRefers to itself, its category, or the conceptual structures that define it.
23TemporalTime plays a defining role in identity, structure, or effect.
24Digital/VirtualExists primarily in digital form, as data, software, or virtual representation.
Edge cases — Abstract Layer
1Wedding ring — Symbolizes marriage/commitment beyond material form
0Abstract art — Doesn’t represent something else; is itself
1Smoke signal — Intentionally transmits information
0Smoke from fire — Not intentionally communicating
1Crystal structure — Follows precise geometric/chemical rules
0Freestyle dance — No formal rules govern movements
1Dictionary — Language describing language (meta)
0Regular novel — First-order narrative, not self-referential
1Clock — Inherently about measuring/displaying time
0Photograph — Captures a moment but is itself static

BitTraitDescription
25Social ConstructExists by virtue of collective human understanding or cultural meaning.
26Institutionally DefinedDefinition depends on formal institutional rules, standards, or designations.
27Identity-LinkedTied to personal, social, or occupational identity, roles, or status.
28RegulatedSubject to formal or informal rules, laws, or institutional oversight.
29Economically SignificantHas measurable economic value, cost, or role in economic transactions.
30PoliticisedSubject to political discourse, debate, or partisan disagreement.
31RitualisedAssociated with ceremonies, traditions, or culturally significant observances.
32Ethically SignificantRaises ethical questions or involves moral considerations.
Edge cases — Social Layer
1Money (physical bill) — Value is socially constructed; paper is just the medium
0Mountain — Physical formation, not social construction
1Endangered species (label) — Designation is institutional (legal protection status)
0Species (biological) — Natural category defined by biology, not institutions
1Chef’s hat — Signifies chef role/identity
0Generic hat — No identity association
1Champagne — Associated with celebrations and toasts (ritualized)
0Beer (generic) — Everyday beverage, not inherently ritualistic
1AI system — Bias, privacy, and autonomy concerns are ethical issues
0Spoon — Morally neutral object

04 Classification Methodology

UHT classification uses a strict, minimalist approach. Each of the 32 traits is evaluated independently against the entity under consideration.

Inclusion Standard: A trait should be marked Included (1) only if it is intrinsically present in the entity’s identity. The trait must be:

  • Structurally inseparable from the entity
  • Functionally designed-in (part of its purpose or operation)
  • Logically entailed by its construction or symbolic role

Exclusion Rule: A trait should be marked Excluded (0) if it is merely attributed in context, present through metaphor, or not essential to the entity’s core identity.

Avoid Trait Inflation: Do not include traits based on secondary usage, surrounding systems, or cultural interpretation — unless that role is intrinsically embedded in the entity’s identity (e.g., a “Police Badge” is intrinsically symbolic and social).

Justification Requirement

For each of the 32 traits, the classifier must indicate whether it is Included or Excluded and provide a concise justification based on the official trait definitions. This ensures auditability and reproducibility across classifiers.

Automated Classification Pipeline

The UHT Factory uses an LLM-driven classification engine:

  1. Cache check — Redis lookup by entity name
  2. Parallel trait evaluation — 32 async tasks sent to an LLM, each evaluating one trait
  3. UHT code construction — 32-bit binary string converted to 8-character hex
  4. Storage — Redis cache + Neo4j graph database

Each trait evaluation produces a JSON object with applicable (boolean), confidence (0.0–1.0), and justification (max 100 words).

05 Encoding Guide

Encoding Workflow

  1. Evaluate all 32 traits using the official definitions
  2. Justify every inclusion/exclusion
  3. Build the 32-bit binary string (MSB = bit 1)
  4. Group into four 8-bit bytes (one per layer)
  5. Convert each byte to a 2-character hex pair
  6. Concatenate: [PHYS][FUNC][ABST][SOC]

Decoding Workflow

  1. Split the 8-character hex code into four 2-character pairs
  2. Convert each pair to an 8-bit binary string
  3. Map each bit to its corresponding trait name
  4. Use the active traits for search, comparison, or explanation

Bit Ordering

Within each layer byte, Bit 1 (of that layer) corresponds to the Most Significant Bit (MSB, position 27). Bit 8 maps to the Least Significant Bit (LSB, position 20). This means the first trait in each layer controls the leftmost bit of that hex pair.

Output Format

{ "entity": "Chess", "uht_code": "4090FDCF", "traits": { "physical": [], "functional": ["Intentionally Designed"], "abstract": ["Symbolic", "Rule-governed", "Compositional", ...], "social": ["Social Construct", "Regulated", ...] }, "feedback": { "Physical Object": "Excluded — Chess is a game, not a physical object", // ... 32 justifications } }

06 Semantic Compression

UHT codes act as a form of semantic compression, reducing complex identity into 32 compact, interpretable bits. This encoding captures the essential meaning of an object or concept into a fixed-size representation.

Benefits

  • Compactness — Each code is only 32 bits (8 hex characters), enabling efficient storage
  • Speed — Bitwise operations like XOR and Hamming distance allow for fast comparisons
  • Clarity — Each bit is linked to a clear trait, making encodings interpretable and auditable

Semantic Proximity via Hamming Distance

Comparing two UHT codes by their Hamming distance — the number of differing bits — reveals how semantically similar two entities are. The more bits they share, the closer they are in meaning.

Entity ACodeEntity BCodeDistance
Mechanical ClockDEF42205 Digital Alarm ClockDEF82205 2
Coffee MachineDEFA2205 Water Filtration UnitDFD60205 5
PaperclipD78800C5 National Anthem0080C9FF 18

Unlike neural embeddings or black-box classification systems, UHT codes remain human-readable. Every trait is explicit. Differences between entities can be directly traced to the bits that differ.

07 Worked Examples

Paperclip
D78800C5
11010111100010000000000011000101
Physical Object Synthetic Structural Observable Physical Medium Intentionally Designed Human-Interactive

A manufactured object that holds papers by shape. Recognized universally as a basic office tool.

Smart Thermometer
EEFC2205
11101110111111000010001000000101
Physical Object Synthetic Biological/Biomimetic Observable Physical Medium Intentionally Designed Outputs Effect Processes Signals/Logic State-Transforming Human-Interactive System-Integrated Compositional Temporal

A modern sensor that interprets and reports temperature in regulated systems like healthcare and HVAC.

Digital Alarm Clock
DEF82205
11011110111110000010001000000101
Physical Object Synthetic Powered Structural Observable Physical Medium Intentionally Designed Outputs Effect Processes Signals/Logic State-Transforming Human-Interactive Compositional Temporal

A timed signaling device for home or institutional use. Emits sound or light to mark events, widely understood in global culture.

08 Applications

UHT codes apply across diverse domains wherever structured comparison, discovery, or metadata is needed.

Aerospace
Tag components like landing gear, avionics, and hydraulics with semantic fingerprints for compatibility checks, traceability in retrofits, and supplier consistency.
Healthcare AI
UHT traits serve as transparent, interpretable features for diagnostic models. Traits like “biological,” “system-critical,” and “regulated” clarify what a model is interpreting.
Financial Data Integration
A shared semantic layer for tagging transactions, agreements, and compliance records across siloed systems. Enables unified search and audit.
Education
Tag lessons, assessments, and learning objectives by cognitive function — logical, symbolic, behavior-guiding, or socially constructed — to design coherent pathways.
Game Design
Encode characters, items, and rituals. Procedural generation can use trait combinations to create culturally consistent, internally coherent worlds.
Search & Discovery
Trait-based discovery for repositories and knowledge graphs. Filter for “temporal,” “regulated,” or “part of a system” regardless of format.
Supply Chain
Semantic mapping of goods, facilities, and paperwork. Social traits like “regulated” and “economically significant” enable automatic compliance checks.

09 API Reference

The UHT API is served at factory.universalhex.org/api/v1/. Interactive Swagger documentation is available at factory.universalhex.org/docs.

Entities
GET /api/v1/entities List entities with pagination, search, and filtering
GET /api/v1/entities/{uuid} Get entity details with full trait evaluations
POST /api/v1/entities/agent-classify Submit pre-computed classifications
Classification
POST /api/v1/classify Classify an entity via LLM (32 parallel trait evaluations)
GET /api/v1/classify/status/{task_id} Check classification task progress
Traits
GET /api/v1/traits Get all 32 canonical trait definitions grouped by layer
GET /api/v1/traits/layer/{name} Get traits for a specific layer (Physical, Functional, Abstract, Social)
GET /api/v1/traits/meta-classes Named archetypes for frequent hex pairs
GET /api/v1/traits/prompts Full classifier prompts for each trait
Dictionary
GET /api/v1/dictionary/words Search and browse polysemous words
GET /api/v1/dictionary/words/{lemma} Get word senses with UHT classifications
Graph & Analytics
GET /api/v1/graph/neighbors/{uuid} Find entities connected in the taxonomy graph
GET /api/v1/analytics/trait-cooccurrence Trait co-occurrence matrix across all entities
GET /api/v1/explorer/embeddings UMAP/PaCMAP projections for entity similarity visualization
Hex Calculator
GET /api/v1/hex-calc/xor XOR two UHT codes to find differing traits
GET /api/v1/hex-calc/hamming Compute Hamming distance between two codes

10 Comparison to Existing Approaches

UHT can be compared to several established methods for representing identity, meaning, or metadata.

MethodCommon UsageLimitationsUHT Advantage
RDF / OWL Ontologies Semantic web, linked data Schema-heavy, domain-specific One format usable across all domains
Neural Embeddings Machine learning, NLP Opaque, unstable, hard to explain Transparent, fixed-length, human-readable
ISO Metadata Standards Documentation, archiving Rigid, verbose, hard to compare Compact, trait-based, easy to cluster
Feature Flags / Tags Software configs, taxonomies Flat, unstructured, lacks depth Layered structure and formal bit encoding

UHT stands out by offering a hybrid: a fixed-length, interpretable code that is small enough to fit anywhere but expressive enough to work across physical, abstract, and social contexts.