Intellectual and Technical Foundations of the Embedded Semantics Architecture

Full report

On this page

1. Executive Summary

The Embedded Semantics architecture introduces a foundational paradigm shift in knowledge representation and artificial intelligence communication: natural-language expressions and visual glyphs are interpreted strictly as evidence regarding meaning, rather than as stable semantic identities. This principle addresses a critical vulnerability in contemporary computational linguistics and large language models (LLMs), which frequently conflate the surface rendering of text or its statistical distributional vector with its underlying conceptual reference. By architecting a deliberate, non-permeable boundary between the visible expression and the inferred concept, the system provides a mathematically and philosophically robust framework for semantic retrieval, multilingual alignment, and interoperable machine reasoning. The operational pipeline follows a distinct progression. It begins with the Expression (the surface form), proceeds through semantic candidate discovery utilizing vector embeddings, and culminates in a stable, registry-backed Concept Identity. This Concept Identity is inherently versioned, inspectable, and capable of retaining complex metadata, including provenance, definitions, structural relationships, and quantifiable uncertainty1. Crucially, while embeddings and vector coordinates are leveraged for probabilistic discovery and alignment, they are explicitly prohibited from serving as the authoritative semantic identity. This methodology guarantees that geometric similarity does not silently overwrite logical identity, thereby supporting true semantic convergence, explicit ambiguity management, and machine-readable provenance. The ensuing analysis meticulously examines the historical, linguistic, philosophical, and computational foundations of this architecture, demonstrating its alignment with established intellectual traditions while highlighting its structural advantages over opaque generative systems.

2. Embedded Semantics Thesis

The core thesis of the Embedded Semantics architecture posits that a linguistic expression—whether a sequence of Unicode characters, a synthesized voice command, or a rendered glyph—is merely an artifact of communication. It is a transport mechanism, not the locus of meaning itself. The architecture actively resists the collapse of the signifier and the signified into a single computational unit, a flaw prevalent in systems that rely exclusively on end-to-end vector mappings. To operationalize this thesis, the architecture deploys a multi-layered pipeline, best exemplified by the project's Four-Layer Glyph Object Specification: Surface, Structure, Embedding, and Canonical layers1. The Surface layer captures the visible expression, managing Unicode sequences, normalization policies, and rendering profiles. The Structure layer maps the visible mechanics, such as syntactic relations or visual components, without assigning meaning. The Embedding layer utilizes high-dimensional vectors to gather evidence across visual, structural, and semantic lanes. Finally, the Canonical layer resolves this evidence into a bounded gloss and a stable Concept Identity, accompanied by a calibrated confidence score and explicit warnings1. By treating the natural language expression purely as an evidentiary input, the system can execute a "no-op" (doing nothing) when evidence is insufficient, refusing to hallucinate a false mapping1. This creates an architecture where the visible mechanics of communication remain distinctly separate from the authoritative meaning, allowing the system to handle explicit ambiguity, trigger human review, and seamlessly manage multilingual semantic convergence.

3. Historical Foundations

The architectural separation of expression from meaning is deeply rooted in semiotics, the philosophical and historical study of signs and symbols. The foundational work of Ferdinand de Saussure established the dyadic model of the sign, bifurcated into the signifier (the physical form of the sign, such as a written word or sound wave) and the signified (the mental concept it represents). The Embedded Semantics architecture essentially codifies this semiotic divide into its computational infrastructure. By treating natural language expressions solely as evidence, the architecture acknowledges Saussure's principle of the arbitrariness of the sign: the physical form has no inherent logical connection to the concept it conveys; it is entirely governed by fluid cultural and linguistic convention. Charles Sanders Peirce expanded this framework into a triadic model, introducing the "referent" (the actual object or truth in the world to which the sign refers), alongside the representamen (the form) and the interpretant (the sense made of the sign). In the context of the Embedded Semantics ecosystem, the surface expression aligns with the representamen. The candidate discovery and vector embedding layers function as the interpretant, mapping the physical form to a cognitive space. The stable, registry-backed Concept ID serves as the formal surrogate for the referent within the system's ontology. By preventing vector embeddings from serving as the ultimate identity, the architecture avoids the semiotic error of mistaking the interpretant for the referent, ensuring that the system's internal reasoning remains anchored to explicitly defined, versioned parameters.

4. Linguistic Foundations

The linguistic underpinnings of the Embedded Semantics framework are heavily informed by advancements in lexical semantics and semantic field theory, contrasted sharply against the limitations of pure distributional semantics. Distributional semantics, summarized by J.R. Firth's adage that a word is characterized by the company it keeps, forms the theoretical basis of modern vector embeddings. While this theory effectively captures relatedness and usage patterns, it inherently fails to establish stable semantic boundaries. Distributional models excel at placing words in a continuous latent space but struggle with antonymy, hypernymy, and explicit polysemy, as diametrically opposed concepts (e.g., "always" and "never") often appear in identical syntactic contexts. Embedded Semantics addresses this deficiency by integrating principles from prototype theory and conceptual spaces. Prototype theory, pioneered by Eleanor Rosch, suggests that concepts are organized around central, prototypical examples rather than rigid necessary and sufficient conditions. The architecture's inclusion of positive and negative examples within the Concept ID registry mirrors this cognitive organization, allowing for graded categorization while maintaining a fixed identifier for the category itself. Furthermore, conceptual spaces, as theorized by Peter Gärdenfors, provide a geometric representation of knowledge where concepts are regions within quality dimensions. While vector embeddings approximate this, they lack the rigid topological boundaries required for logical operations. Semantic field theory posits that words gain meaning through their structural relationships with other words in a specific domain. The Embedded Semantics architecture operationalizes these semantic fields by explicitly encoding relationships, context, and structural graph analyses within the registry, ensuring that compositional meaning is derived from structured, inspectable relationships rather than opaque vector arithmetic1.

5. Philosophy of Meaning

The most profound philosophical foundation for the Embedded Semantics architecture is found in the work of Gottlob Frege, specifically his 1892 treatise Über Sinn und Bedeutung (On Sense and Reference)2. Frege identified a critical paradox in identity statements. He noted that the statement [Figure omitted from source export] is an uninformative tautology, while [Figure omitted from source export] contains cognitive value, provided it is true4. Using the classical astronomical example, both the "Morning Star" and the "Evening Star" refer to the same physical entity (the planet Venus), meaning they share the same reference (Bedeutung). However, they carry different cognitive significance or modes of presentation, which Frege termed sense (Sinn)3. Embedded Semantics computationally formalizes Frege's distinction to resolve the modern artificial intelligence hallucination crisis. The natural language expression and its subsequent vector embedding correspond to the Sinn—the mode of presentation, heavily influenced by context, language, and cultural syntax. The stable, registry-backed Concept ID corresponds to the Bedeutung—the authoritative, unshifting reference point. If a system relies solely on language models and vector spaces, it effectively collapses reference into sense, attempting to extract stable truth from subjective modes of presentation. This relates directly to the distinction between intensional and extensional meaning. Extensional meaning is the set of entities in the world to which a concept applies, while intensional meaning is the internal logical definition or properties that determine those entities. Vector spaces are largely extensional, derived from statistical sets of word co-occurrences. The registry-backed Concept ID provides the rigorous intensional rule. Furthermore, the architecture incorporates Ludwig Wittgenstein’s later philosophy of "meaning as use." By treating expressions as evidence and explicitly tracking human comprehension and cohort effects, the system acknowledges that words operate as tools within language games, requiring explicit mapping of ambiguity and pragmatic context rather than assuming a universal, hidden lexicon1.

6. Computational Semantics

To operationalize these philosophical and linguistic theories, the architecture bridges the historical divide between Symbolic AI and modern statistical Natural Language Processing (NLP). Symbolic AI relied heavily on explicit knowledge representation schemas, such as semantic networks and semantic primitives. While highly interpretable and precise, these systems suffered from a knowledge acquisition bottleneck and catastrophic brittleness when faced with natural language variation. Attempts to bridge this gap included Controlled Natural Languages (CNLs). CNLs, such as Attempto Controlled English (ACE), are engineered subsets of natural languages designed to eliminate ambiguity7. ACE restricts lexicon and syntax to allow deterministic, unambiguous translation into formal logic representations, specifically Discourse Representation Structures (DRS), enabling automated reasoning9. While CNLs provide perfect semantic mapping, they enforce unnatural, rigid constraints on the human user8. Embedded Semantics adopts the logical rigor of CNLs but relocates the rigidity from the user interface to the internal registry. The user is permitted to use fully unstructured, natural expressions. The system leverages modern computational semantics—specifically vector embeddings and lexical databases like WordNet (which groups words into synonym sets or "synsets")—for the "soft" task of candidate discovery, absorbing the variance of natural language. However, the final output resolves to a strict, symbolic Concept ID linked via Frame Semantics. Frame semantics, formulated by Charles Fillmore, posits that meaning is understood as a structure of interconnected roles11. The Canonical Layer of the architecture assembles these IDs into explicit, machine-readable packets where roles and relations are preserved, successfully merging statistical flexibility with symbolic determinism1.

7. Vector Representations

Vector representations and latent semantic spaces represent the current state-of-the-art in semantic candidate discovery. By mapping discrete text tokens into continuous, high-dimensional spaces, geometric distance can be used as a proxy for semantic similarity. While embeddings are highly efficacious for retrieval algorithms and establishing initial multilingual alignments, their underlying architecture poses an existential risk if mistaken for semantic identity. Embeddings are mathematically fluid. They change entirely with model retraining, fine-tuning, and natural shifts in the underlying training corpus. A set of floating-point coordinates in a 1,536-dimensional space cannot serve as a stable, cross-system semantic identifier because the space itself is not persistent. Furthermore, neural embeddings suffer from the "Vector Illusion"—the assumption that high-dimensional proximity guarantees logical substitutability. LLM latent spaces excel at pattern matching but are entirely devoid of an explicit ontology. When an architecture relies exclusively on a latent space, it abandons the ability to perform precise concept normalization, making it impossible to guarantee semantic interoperability across distinct systems or organizations. Therefore, the Embedded Semantics framework dictates that while vectors are essential for evidence gathering and candidate ranking, they must terminate at a rigid registry boundary before any authoritative claim is made1.

8. Symbolic and Registry-Based Representations

The necessity of a stable, registry-backed ontology is best evidenced by large-scale medical and terminology systems, which serve as the gold standard for separating text from concept. The Unified Medical Language System (UMLS) functions as a comprehensive metathesaurus, providing a mapping structure among diverse medical vocabularies and linking disparate textual names (synonyms, acronyms, translations) to a single, normalized concept12. Similarly, the Systematized Nomenclature of Medicine Clinical Terms (SNOMED CT) operationalizes this by organizing clinical terms into acyclic taxonomic hierarchies13. Each SNOMED CT concept possesses a unique numerical identifier and a Fully Specified Name (FSN) that includes a semantic tag to guarantee disambiguation13. This architecture explicitly recognizes that multiple linguistic expressions are merely evidence pointing to a single, invariant medical reality. The Embedded Semantics architecture scales this medical-grade rigor to general-purpose language and semantic glyph interpretation. By maintaining stable boundaries, definitions, positive/negative examples, and relationships outside of the retrieval mechanism, the architecture allows for explicit abstention. If an expression does not confidently resolve to a known Concept ID, the system flags an error or triggers a human review queue rather than hallucinating an approximate response—a critical safety feature absent in pure generative AI architectures1.

9. Why Similarity Is Not Identity

The most severe flaw in contemporary vector-centric architectures is the conflation of geometric similarity with semantic identity. The Embedded Semantics framework mandates a strict demarcation between seven distinct semantic states, ensuring they are not collapsed into a single vector-distance interpretation. Collapsing these into a single scalar value destroys the nuance required for high-stakes reasoning.

The Embedded Semantics architecture uses similarity strictly for discovery, but enforces identity, equivalence, and entailment via the registry constraint, ensuring that geometric proximity never overrides explicit logical definitions.

10. Multilingual Meaning

Modern language models generate impressive translations but fundamentally lack an explicit interlingua—a language-agnostic representation of meaning. When a concept transitions from English to Japanese within an LLM, it traverses a continuous latent space without ever passing through an inspectable, language-neutral checkpoint. This introduces unmanageable risks of cultural and semantic drift, where idioms or cultural biases silently corrupt the translation. Embedded Semantics relies on Concept IDs as the ultimate interlingua. This aligns historically with initiatives like the Universal Networking Language (UNL), which sought to provide a deep semantic representation framework facilitating translation through language-independent nodes and relations, yielding high accuracy in morphology, syntax, and semantics14. In the Embedded Semantics paradigm, the Concept ID is the sovereign entity. English expressions, Mandarin expressions, and standardized visual glyphs (such as those developed in the Protocol5 IOTA-1 workbench) are all mapped as surface representations of this singular ID1. Consequently, multilingual alignment is not a matter of translating word-for-word, nor is it subject to latent space approximations; it is the act of linking diverse linguistic evidence to a stable ontological coordinate. This facilitates true semantic convergence and interoperability.

11. Stable Semantic Identity

For a system to support machine-readable semantic packets, meaning must possess absolute persistence. If the underlying discovery model is retrained and the vector space shifts, the meaning must survive the transition intact. Stable Semantic Identity requires that Concept IDs function as Persistent Identifiers (PIDs). Much like DOIs in academic publishing or UUIDs in database architecture, these semantic identities must be immutable. If the definition of a concept evolves over time, the architecture requires the Semantic Versioning of Knowledge. A Concept ID is not overwritten; a new version is minted, and explicitly linked to the predecessor, detailing the precise nature of the conceptual drift. This allows archival data to remain historically accurate while enabling forward-compatibility. The registry acts as the sole source of truth, heavily curated and version-controlled, providing a bedrock that volatile neural network weights cannot offer. The ecosystem relies heavily on UAIX protocols and Agent File Handoff mechanisms to ensure these identities are portable, auditable, and contractually binding across different platforms16.

12. Compositional Meaning

A critical vulnerability of sentence-level vector embeddings is their inability to retain explicit structural composition. When an entire sentence is compressed into a single vector, the roles of the individual agents, actions, and objects are blended. Recovering the exact relationship from this vector is a mathematically lossy process. Embedded Semantics addresses compositional meaning by drawing on Vector Symbolic Architectures (VSA), also known as Hyperdimensional Computing (HDC)18. VSAs utilize extremely high-dimensional vectors (often 10,000 dimensions or more) and an algebra of operations—specifically binding, bundling (superposition), and permutation—to represent complex, compositional data structures20. Unlike traditional neural network embeddings, which lose structural hierarchy when compressing data, VSAs can explicitly encode trees, graphs, and sequences within a single vector without expanding the dimensionality21. However, while VSAs attempt to encode the symbolic structure inside the mathematics of the vector itself, Embedded Semantics utilizes this principle for the Structure Layer but enforces a rigid structural separation at the Canonical Layer1. It mandates that the final resolution occurs at the registry level, explicitly linking discrete Concept IDs. This mirrors the representational power of VSAs but avoids the combinatorially hard search spaces associated with decomposing complex hypervectors, ensuring that composition remains fully inspectable, logically verifiable, and immune to the combinatorial explosions that plagued traditional Symbolic AI19.

13. Provenance and Versioning

Data without provenance is computationally hazardous. The World Wide Web Consortium (W3C) established the PROV Data Model (PROV-DM) to provide a conceptual framework for inter-operable provenance interchange on the web, distinguishing core structures of provenance from extended use cases23. Embedded Semantics adopts the exact spirit of PROV-DM by mandating that every resolved Concept ID carries a permanent trace of its origin. The system's Teleodynamic State explicitly tracks the resource budget, maintenance burden, and the "phase regime" of the interpretation1. When an expression maps to a Concept ID, the system records the vector distances, the structural graph analysis, the specific version of the registry used, and the calibrated confidence score1. Crucially, if the interpretation requires a "split", "merge", or a "no-op" due to insufficient evidence, this structural edit is heavily logged. This provenance data guarantees that downstream systems, human reviewers, or adjacent ecosystem components (such as the ErrorNotifier.com telemetry lane or the AIWikis.org long-memory storage) can audit exactly why a specific expression was resolved to a specific Concept ID17. This is an essential feature for safety-critical applications, regulatory compliance, and cognitive liberty protocols.

14. Major Counterarguments

An exhaustive architectural design must confront and analyze its strongest criticisms. The following outlines valid counterarguments to the Embedded Semantics philosophy and articulates how the project mitigates them or utilizes them as intentional boundaries.

15. Where Embedded Semantics Is Strong

The architecture excels in environments demanding auditability, cross-system interoperability, and long-term semantic stability. By divorcing the expression from the identity, the system is immune to linguistic drift and model retraining. It possesses a distinct advantage in explicitly modeling ambiguity; whereas an LLM will confidently hallucinate an answer when faced with an ambiguous prompt, the Embedded Semantics framework triggers a deterministic failure, records the semantic residue, or routes to a human review queue. Furthermore, the architecture is supremely positioned for the exact goals of the JustAnIota.com and Protocol5.com ecosystems1. By utilizing the IOTA-1 (ɪ≃1) approximate interpretation protocols, the system allows for compact public-symbol semantic conversion that is evidence-first, rather than relying on hidden codebooks or claiming exact, lossless translation1. The integration of UAI-1 schemas ensures that AI memory packages and handoffs remain strictly governed by portable evidence16.

16. Where the Project Must Remain Humble

The project must strictly demarcate its capabilities to avoid the hyperbolic pitfalls common in modern artificial intelligence discourse. The Teleodynamic.com philosophical fulcrum explicitly mandates that the architecture is a structural framework for communication; it is not, and does not claim to be, Artificial General Intelligence (AGI)24. The existence of a robust, registry-backed ontology does not imply that the system possesses intrinsic understanding, consciousness, sentience, or biological equivalence27. Furthermore, the system does not own "universal truth." The Concept IDs are operational definitions used to facilitate interoperability, not philosophically absolute categorizations of the universe. The project maintains explicit humility regarding human review: the architecture operates on the principle that machine comprehension is merely a probabilistic proxy for human comprehension, and human oversight remains the ultimate arbiter of semantic validity. The system's true strength lies in its ability to securely manage its own ignorance through the "no-op" function, acting only when evidence surpasses rigorous, predefined thresholds1.

17. Terminology Recommendations

To maintain theoretical rigor and prevent conceptual drift across EmbeddedSemantics.com, UAIX.org, Protocol5.com, and JustAnIota.com, the following precise terminology should be enforced universally:

18. Site Content Opportunities

EmbeddedSemantics.com possesses a unique opportunity to position itself alongside Teleodynamic.com as a definitive intellectual hub for verifiable AI communication. Current documentation should be expanded to include:

1. The Registry Governance Model: A transparent guide detailing exactly how Concept IDs are minted, versioned, and deprecated. This directly addresses the "bureaucratic rigidity" counterargument by showing lightweight governance in action.

2. The LLM Integration Handbook: Clear technical documentation for developers showing how to use LLMs purely for Semantic Candidate Discovery without letting generative hallucinations pollute the Canonical Concept layer.

3. Cross-Disciplinary Glossary: A dedicated page mapping Embedded Semantics terminology directly to historical counterparts in Semiotics (Signifier/Signified), Philosophy (Sinn/Bedeutung), Medical Terminology (SNOMED FSNs), and Computational Linguistics (VSAs/Ontologies).

4. Failure Mode Matrix: A public ledger of known architectural limitations and how the "no-op" protocol handles them, reinforcing the project's commitment to the boundaries defined by the philosophical fulcrum27.

19. Proposed Cornerstone Articles

To educate the target audience, establish intellectual authority, and address persistent misconceptions, the following cornerstone articles must be developed for EmbeddedSemantics.com.

20. Recommended Diagrams

Visualizing this complex architecture is paramount for user comprehension. The following diagrams should be engineered and integrated into the public documentation:

1. The Semiotic Pipeline Map: A flow diagram illustrating an input expression moving left-to-right. It passes through the Surface Layer (normalization), splits into vector coordinates in the Embedding Layer (labeled explicitly as Sense / Mode of Presentation), narrows through a filter labeled "Heuristics/Ontology Checks", and locks into a rigid, central block labeled Canonical Concept ID (labeled as Reference). The output branches out to machine-readable semantic packets.

2. Vector Proximity vs. Ontological Distance: A comparative side-by-side graphic. On the left, a 3D scatter plot showing words like "Increase" and "Decrease" clustered closely together (Vector Space). On the right, a structured graph showing the exact same words on opposite ends of a defined semantic field, explicitly mediated by an antonymy edge (Embedded Semantics Registry).

3. The Four-Layer Object Stack: A vertical stack diagram illustrating a single semantic transaction to visualize the glyph object specification1.

• Top Layer: Surface (e.g., Input string: ɪ≃1)
• Second Layer: Structure (Visual mechanics / Syntax roles)
• Third Layer: Embedding (Evidence vectors, visual, semantic)
• Bottom Layer: Canonical (Concept ID, Bounded Gloss, Confidence \= 0.68, Status \= Emerging)

21. Research Gaps

While the foundations of Embedded Semantics are exceptionally robust, several critical areas require further academic and technical exploration to ensure long-term viability:

• Dynamic Contextual Binding: While Vector Symbolic Architectures (VSAs) offer theoretical mechanisms for binding context to concepts, computationally efficient methods for mapping fluid vector context (e.g., irony, sarcasm, idiom) to a rigid Concept ID without generating an unmanageable explosion of sub-variants remains challenging.
• Automated Ontology Alignment: Discovering how to securely map newly discovered concepts against existing, massive registries (like UMLS or SNOMED CT) in real-time, without violating the strict "no-op" safety principle, requires extensive empirical testing and bridging software.
• Quantifying Semantic Residue: Developing standard mathematical metrics for measuring exactly how much meaning is "lost" or left behind when a complex expression is compressed into a single Concept ID. Identifying the mathematical threshold of residue at which an ID assignment should automatically fail is critical for future versions.
• Evolution of the Interlingua: Investigating the long-term governance of Concept IDs to prevent Anglocentric or domain-specific cultural biases from silently dominating the underlying ontology structures over time.

22. Conclusions

The Embedded Semantics architecture successfully navigates the historical limitations of both brittle symbolic artificial intelligence and fluid, unanchored generative language models. By foundationalizing the premise that natural-language expressions and visual glyphs are merely evidence about meaning, the architecture secures a mathematically and philosophically sound methodology for semantic processing. Rooted in the semiotic traditions of Saussure and Peirce, and strictly adhering to the logical rigor of Frege’s separation of sense and reference, the system recognizes that the physical signifier (the text or embedding) and the psychological signified (the concept) must be computationally decoupled. It leverages the massive retrieval power of modern vector spaces and the structural integrity of Hyperdimensional Computing to handle the chaos of natural language. Concurrently, it relies entirely on explicit, registry-backed Concept IDs to guarantee semantic identity, interoperability, and truth-functional entailment. The architecture’s unwavering commitment to provenance tracking, semantic versioning, explicit compositional structures, and the "no-op" abstention protocol renders it exceptionally suitable for high-stakes, multilingual, and cross-domain applications. By refusing to collapse geometric similarity into logical identity, Embedded Semantics offers a paradigm where artificial intelligence can achieve verifiable, stable, and deeply auditable knowledge representation, firmly anchoring the volatile surface of human communication to the stable bedrock of structured, mathematical meaning.

23. Annotated Source List

In accordance with the project requirements, the following annotated framework maps the utilized research data and external theories to the core architecture of Embedded Semantics:

• Teleodynamic Ecosystem & Bounded Claims: Source materials outlining the philosophical fulcrum, claim boundaries, and the "no-op" discipline of Teleodynamic AI, establishing the project's refusal to claim AGI or consciousness24.
• UAIX Schema & Handoff Protocols: Documentation on UAI-1 schema ownership, portable evidence lanes, and memory hygiene, which provide the structural rules for how Concept IDs are passed between agents16.
• Protocol5 & JustAnIota (IOTA-1): Source materials defining the approximate interpretation (ɪ≃1) workbench, Unicode governance, and the four-layer glyph object specification (Surface, Structure, Embedding, Canonical) that separates visual expression from internal semantics1.
• Fregean Philosophy of Language: Foundational philosophical texts regarding Gottlob Frege’s Über Sinn und Bedeutung, used to establish the architectural necessity of separating mode of presentation (Sense) from registry identity (Reference)2.
• Vector Symbolic Architectures (VSA) / Hyperdimensional Computing: Academic surveys on computing in superposition, illustrating how VSAs combine distributed representations with symbolic structure without losing dimensionality18.
• Medical Knowledge Registries (SNOMED CT, UMLS): Technical documentation on the Unified Medical Language System and SNOMED CT, providing the real-world standard for mapping varied textual expressions (synonyms) to stable, multi-axial, persistent Concept IDs12.
• Controlled Natural Languages (CNL): Surveys on systems like Attempto Controlled English (ACE), demonstrating the historical drive to translate ambiguous English directly into formal first-order logic and Discourse Representation Structures7.

Works cited

1. Semantic Glyph Systems and Teleodynamic AI Communication, https://teleodynamic.com/glyph-communication/

2. Sinn/Bedeutung \- Oxford Reference, https://www.oxfordreference.com/display/10.1093/oi/authority.20110803100508390

3. Sense and reference \- Wikipedia, https://en.wikipedia.org/wiki/Sense\_and\_reference

4. From The Begriffsschrift To "Über Sinn Und Bedeutung": Frege As Epistemologist And Ontologist \- SciELO, https://www.scielo.br/j/man/a/R3mRbHSGtDgzWhXLGjhmXgG/?lang=en

5. (PDF) From The Begriffsschrift To "Über Sinn Und Bedeutung": Frege As Epistemologist And Ontologist \- ResearchGate, https://www.researchgate.net/publication/311238972\_From\_The\_Begriffsschrift\_To\_Uber\_Sinn\_Und\_Bedeutung\_Frege\_As\_Epistemologist\_And\_Ontologist

6. Bedeutung l, http://fitelson.org/proseminar/frege\_osar.pdf

7. A Survey and Classification of Controlled Natural Languages \- ACL Anthology, https://aclanthology.org/J14-1005.pdf

8. Controlled natural language \- Wikipedia, https://en.wikipedia.org/wiki/Controlled\_natural\_language

9. (PDF) Attempto Controlled English for Knowledge Representation \- ResearchGate, https://www.researchgate.net/publication/225162366\_Attempto\_Controlled\_English\_for\_Knowledge\_Representation

10. Controlled Natural Languages for Knowledge Representation \- ACL Anthology, https://aclanthology.org/C10-2128.pdf

11. Frame Semantics \- Brill \- Reference Works, https://referenceworks.brill.com/display/entries/HCSO/COM-0106.xml?language=en

12. Unified Medical Language System \- Wikipedia, https://en.wikipedia.org/wiki/Unified\_Medical\_Language\_System

13. SNOMED CT \- Wikipedia, https://en.wikipedia.org/wiki/SNOMED\_CT

14. UNL in Machine Translation Systems | PDF | Semantics \- Scribd, https://www.scribd.com/document/872056133/unl-nlp

15. The Universal Networking Language beyond Machine Translation, https://www.semanticscholar.org/paper/The-Universal-Networking-Language-beyond-Machine-Uchida/a54811b8d33de92098761c96ca944cd083587e78

16. Contact / About Michael Kappel \- Cognivirus.com, https://cognivirus.com/contact/

17. Ecosystem overlay and domain authority boundaries \- Teleodynamic AI, https://teleodynamic.com/ecosystem-overlay/

18. Vector Symbolic Architectures as a Computing Framework for Emerging Hardware \- arXiv, https://arxiv.org/abs/2106.05268

19. Self-Attention Based Semantic Decomposition in Vector Symbolic Architectures \- arXiv, https://arxiv.org/html/2403.13218v1

20. \[2111.06077\] A Survey on Hyperdimensional Computing aka Vector Symbolic Architectures, Part I: Models and Data Transformations \- arXiv, https://arxiv.org/abs/2111.06077

21. hyperdimensional computing: a fast, robust and \- arXiv, https://arxiv.org/pdf/2402.17572

22. Residual and Attentional Architectures for Vector-Symbols \- arXiv, https://arxiv.org/pdf/2207.08953

23. PROV-DM: The PROV Data Model \- W3C, https://www.w3.org/TR/2012/CR-prov-dm-20121211/diff.html

24. Teleodynamic.com Is the Philosophical Fulcrum of the Ecosystem, https://teleodynamic.com/philosophical-fulcrum/

25. https://teleodynamic.com/unicode-governance/

26. Teleodynamic AI Summary for Machine Readers, https://teleodynamic.com/ai-summary/

27. Teleodynamic AI FAQ and Claim Boundaries, https://teleodynamic.com/claim-boundary-faq/

28. Ecosystem Role Map \- Teleodynamic AI, https://teleodynamic.com/ecosystem-role-map/

29. A Survey on Hyperdimensional Computing aka Vector Symbolic Architectures, Part I: Models and Data Transformations \- arXiv, https://arxiv.org/pdf/2111.06077

30. Systematized Nomenclature of Medicine \- Wikipedia, https://en.wikipedia.org/wiki/Systematized\_Nomenclature\_of\_Medicine