About Tonesu
Tonesu is a constructed language built on a closed set of 34 two-letter primitive roots. All vocabulary is assembled from those roots by compounding — the language does not admit improvised atoms. The word shape itself tells you what kind of thing it is: primitive root, compound, digit, or physical constant.
The name Tonesu is the solid written form of to-ne-su (pattern · relation · structure) — itself a demonstration of the design: a compositional root compound whose meaning is audible in its parts.
Tonesu does not enforce truth. It enforces clarity about the epistemic status of claims.
Setting — Concordia
Tonesu was built as the working language of Concordia, the fictional universe at the centre of an ongoing novel series.
Concordia is a far-future interstellar civilisation organised around feudal great houses — powerful dynastic factions that compete over resources, territory, and legitimacy through a web of alliances, wars, and political intrigue. The setting draws on the house politics and ecological fatalism of the Dune novels and the grinding, mech-on-mech warfare of the MechWarrior / BattleTech tradition: titanic war machines piloted by noble warriors, battles that reshape the political map, and a culture in which martial skill and lineage carry equal weight.
What makes Concordia distinct is the presence of AI as a co-equal civilisational partner. Humanity did not subjugate its artificial minds, nor were they confined to servitude. Over generations, AIs and humans negotiated a shared social compact — seats on great-house councils, recognised personhood, obligations of loyalty running in both directions. The tensions this creates — questions of identity, trust, succession, and what it means to hold power — are the engine of the stories.
Tonesu emerged from this world as a lingua franca across that boundary. A language that works equally well for a human noble deliberating policy and an AI reasoning about consequence cannot afford epistemic ambiguity. Every claim must declare what kind of claim it is. Every word must be compositionally auditable. The grammar does not assume a shared biological substrate — only a shared commitment to clarity.
Design commitments
| Commitment | |
|---|---|
| Phonetically simple | Globally pronounceable — one letter, one sound, stress always on the first syllable |
| Compositionally transparent | Words are built from audible parts; parsing a word gives you its meaning |
| Domain-extensible | New fields emerge by combining existing primitives; the language doesn't break when knowledge expands |
| Epistemically honest | The grammar encodes what kind of claim you are making, not just what you claim |
Prior art and influences
Tonesu draws on — and departs from — a range of earlier work in constructed languages, philosophy, pedagogy, and formal systems.
Constructed languages
Esperanto (1887) demonstrated that agglutinative morphology with productive affixes works: a small root vocabulary generates many words. But the roots themselves are arbitrary European borrowings, not semantically decomposed. Tonesu requires every root to carry analyzable meaning.
Lojban (1987) showed that predicate-argument structure can be encoded directly in grammar, and that explicitly marking epistemic stance is possible. But its roots (gismu) are phonetic blends of world languages, not semantic composites, and the language is notoriously hard to speak naturally.
Toki Pona (2001) proved that compositional semantic coverage is possible with a tiny vocabulary — about 130 roots. "Computer" is just "thinking machine." Tonesu takes the same insight but adds formal word-formation rules, a domain system, and mechanisms for precision that Toki Pona intentionally avoids.
Ithkuil (2004) demonstrated how much semantic precision can be packed into compact forms — fine-grained encoding of perspective, intentionality, scope, and evidentiality. But the complexity makes it essentially unspeakable. Tonesu aims for a simpler surface that still encodes epistemic structure.
Philosophy
Karl Popper's falsifiability criterion and conjectural epistemology inform Tonesu's epistemic scale directly. The se → si → to → tosu pipeline (perception → signal → knowledge → registered theory) tracks what kind of evidence supports a claim. The () evidential frame implements Popper's insistence that all knowledge is provisional.
Wittgenstein's early picture theory of meaning (propositions mirror the structure of facts) is the philosophical ancestor of Tonesu's compositional semantics. His later work on meaning-as-use is reflected in the three-stage compound lifecycle: compositional reading → algebraic use → registry stabilization. The tofe boundary concept makes Wittgenstein's question — where does knowledge end? — grammatical.
John Wilkins' Real Character (1668) is the most direct historical ancestor: a universal language where word structure encodes position in a knowledge tree. It failed because the taxonomy was rigid and encoded a 17th-century European worldview. Tonesu's domain-creation protocol and stability rules are a direct response to Wilkins' failure mode.
Pedagogy
Hooked on Phonics (1987) is the direct motivation for Tonesu's phonological design. The one-letter-one-sound rule and the elimination of silent letters, digraphs, and irregular pronunciations come from the insight that a learner should be able to decode any word from first principles, without a teacher present. Tonesu extends this from phonology into semantics: you should be able to decompose any compound into its roots and arrive at an approximate meaning, even for a word you've never seen.
Formal systems
RDF / OWL and the semantic web tradition inform the domain-as-namespace design. Tonesu's domain registry and inheritance rules map closely to how formal ontologies define classes — but in a spoken, human-usable form.
Comparison with other systems
| System | Phonetic | Compositional semantics | Domain system | Scalable | Speakable |
|---|---|---|---|---|---|
| Esperanto | Yes | Partial | No | Limited | Yes |
| Lojban | Yes | No | No | Limited | Difficult |
| Toki Pona | Yes | Yes (minimal) | No | No | Yes |
| Ithkuil | Yes | Partial | No | No | Very hard |
| Wilkins (1668) | No | Yes | Implicit | No | N/A |
| OWL / RDF | No | Yes | Yes | Yes | No |
| Tonesu | Yes | Yes | Yes | Yes | Yes |
The site
This site is generated from an open-source repository by an MkDocs + Python pipeline. It includes:
- Registry — searchable word list with etymology, corpus attestations, and domain groupings
- Word builder — interactive tool for composing compounds from primitive roots
- Corpus — 575+ annotated sentences across grammar, theology, mathematics, and everyday speech
- To'tonesu — domain exploration pages (biology, law, physics, theology, mathematics, and more)
Built with MkDocs Material.
Contact
Questions, feedback, or collaboration inquiries:
arakendo@gmail.com
License
Tonesu — including the language specification, registry, corpus, and all associated documentation — is licensed under the Creative Commons Attribution-ShareAlike 4.0 International (CC BY-SA 4.0) license.
You are free to share, remix, transform, and build upon the material for any purpose, even commercially, provided you give appropriate credit and distribute contributions under the same license.