Every word you speak
has cousins across the world.

Search any word or concept. The atlas returns every documented form across language families, scored for structural similarity and traced to shared origins.

400+Corpus Entries

12Language Families

3Analytical Axes

0Contributions

Featured connections

High-confidence cognates — words that share structure across unrelated families.

The core proposition

Words diverge at different rates. Bodily and kinship terms remain near-identical across all documented families. Abstract and cultural terms drift furthest. The path of divergence follows recoverable rules.

Conf = w₁·S_cos(VL₁,VL₂) + w₂·P(C₁→C₂) + w₃·Overlap(PV₁,PV₂)

Explore the framework →

Add your language

Know this word in a language not listed? Submit it — the engine scores it and it enters the review queue.

Language

Word

Meaning / context

Dictionary

Cross-referenced across all documented families. Filter by language or browse the full corpus. Annual print editions published under the PUO imprint.

Contribute to
the corpus

The atlas grows through community knowledge. Submit words, roots, and connections from your language. All entries are scored by the PUO engine and enter a review queue before publication.

Submit a connection

Know a word that connects across families? Start with the concept.

Concept or meaning

Your language

Word in your language

Pronunciation or notes (optional)

Your contributions

Every submission is timestamped and attributed to your account.

No submissions yet.

Corpus statistics

400+Core entries

0Community submissions

5Bantu languages

7Other families

The PUO Framework

A predictive, falsifiable model for mapping cognacy across language families. Three analytical axes form a theorem that generates testable hypotheses about structural relationships.

Axis A

Vowel Signatures

Vowels function as semantic anchors. The AEI core is stable across Bantu roots. O and U nominalise and modulate.

Vowel	Function	Example
A	Base / Action / Root	thusa, dira
E	Subjunctive / Negation	thuse, ga a bue
I	Agentive / Continuous	mothusi, moriti
O	Nominalization	thuso, puo, tiro
U	Progressive / Plural	ubuntu, ma pundu

Axis B

Consonant Rotation

Sound shifts follow probability matrices — not random drift. The D↔T↔R↔L chain is the most documented rotation across all families.

Shift	Example	P
D↔T	dira → tiro	0.70
B↔P↔M	Botho → Motho	0.60
S↔F↔Z	mFana ↔ mSana	0.55
R↔L	Tura ↔ Dula	0.50
G↔K↔H	neGus → Kgosi	0.45

Axis C

Morphological Packaging

Prefix intensity, phoneme inventory, and suffix logic determine how roots are realised and how they map across families.

Feature	Description
PI	Prefix intensity (0–1)
MD	Morphological depth
‑ela	Action for / toward others
S‑Series	Remove S → base root revealed
PIF	Phoneme inventory flags

Confidence Score Formula

Conf = w₁ · S_cos(VL₁,VL₂) + w₂ · P(C₁→C₂) + w₃ · Overlap(PV₁,PV₂)

VL = vowel vector · P(C) = consonant rotation probability · PV = morphological packaging vector

The Deviation Principle

Words diverge at different rates depending on how universal the experience they encode. Bodily and kinship terms — the first things named — show near-zero deviation across all families. Abstract, cultural, and traded concepts show the highest drift.

One origin. Shared base. Rule-governed deviation. Recoverable structure.

Open Research Questions

The framework is explicit about what it still requires.

— Corpus-derived vowel vectors (target: 200–300 cognates)
— Formalized prefix packaging metrics (PI, MD, PIF)
— Extension to Afro-Asiatic and Austronesian families
— Exceptions registry — stress-tests, not weaknesses
— Peer-reviewed publication and institutional validation

Research Portal

For academic institutions, universities, and professional researchers. Partner access opens the full corpus, raw scoring APIs, and collaborative annotation tools.

The PUO corpus and theorem are in active development. Research partners contribute to corpus expansion, provide peer review, and gain full API access to the scoring engine. Institutional partnerships include co-authorship on published findings and curriculum integration support.

Ministries of Education

Curriculum Integration

Pilot programme endorsement, teacher certification pathways, and assessment design. Starting with Botswana — structured for continental adoption.

Universities & Research Institutes

Academic Partnership

Co-develop the corpus, contribute peer review, design joint research programmes. Linguistics, African studies, and computational departments are primary partners.

Technology Partners

ASR & Computational Tools

Development of automatic speech recognition for low-resource African languages — a commercially significant application built directly from the corpus.

Corpus sample — available to research partners

Root	Language	Family	A	E	I	O	U	Notes
THUS-	Setswana	Bantu	thusa	thuse	mothusi	thuso	—	help base root
BUA-	Setswana	Bantu	bua	bue	mmui/sebui	puo	—	B↔P mapping
PUNDA-	OtjiHerero	Bantu	punda	punde	—	(mu)pundo	(ma)pundu	progressive
DIR-	Setswana	Bantu	dira	dire?	dirile	tiro	—	D↔T rotation
DUL-	Tswana/Herero/Eng	Cross-family	dula	dule	—	(bo)dulo	—	dwell (English) cognate
KIT-	Setswana	Bantu	—	ikitse	—	kitso	—	knowledge / self-awareness

Request Research Access

Institutional and academic access by application.

Institution / Organisation

Department / Focus

Research interest

My Profile

G

Guest

No languages set

Member since today · 0 contributions to the corpus

Your linguistic identity

Languages you speak