English builds sentences from separate words. Cheyenne builds them inside a single verb. Subject, object, tense, mode and manner all attach to one stem, and the result is a word that translates as a full English clause.
This is polysynthesis, and it is not rare. It characterises large parts of North America, Siberia and Australia, and it is one of the strongest arguments against generalising about language from European examples alone.
How the verb is assembled
An Algonquian verb has a defined internal structure, with slots filled in a fixed order.
| Slot | Carries |
|---|---|
| Person prefix | who is acting |
| Preverb | tense, manner, direction |
| Initial | the core root |
| Medial | an incorporated noun or shape |
| Final | transitivity and the type of object |
| Suffix | agreement, number, mode |
Because the object can be incorporated directly into the verb, there is frequently no separate noun in the sentence at all. The thing acted upon is inside the word describing the action.
Animacy is a grammatical category, not a description. Algonquian languages divide all nouns into animate and inanimate, and the split is only partly about life — some plants, some natural objects and certain culturally significant items are grammatically animate. Verbs take different forms depending on which category their subject and object belong to, so the distinction has to be learned with every noun.
Obviation: two third persons
Algonquian languages distinguish between two third persons in the same stretch of discourse. One is the proximate — the participant currently in focus. The other is the obviative, marked differently, for any further third person.
| English problem | Algonquian solution |
|---|---|
| "He told him he was late" — who? | proximate and obviative are distinct forms |
| Requires names or rephrasing | tracked grammatically through the passage |
English resolves such ambiguity with names, repetition or context. Cheyenne resolves it in the morphology, and the marking persists across sentences — a discourse-level system European languages simply lack.
The sound system
Cheyenne has an unusually small consonant inventory and marks pitch. Vowels carry high, low or mid tone, and tone is contrastive — words differing only in pitch are different words. The practical orthography marks it with accents.
Vowels can also be voiceless in certain positions, whispered rather than voiced, which is uncommon among the world's languages and difficult for learners to produce naturally.
Why machine translation fails here
| Requirement | Cheyenne situation |
|---|---|
| Large parallel corpora | very little text exists |
| Word-level alignment | one word maps to a whole clause |
| Stable word forms | enormous number of possible forms per stem |
| Ongoing text production | few speakers, little new material |
Statistical and neural systems learn from volume, and the volume does not exist. Worse, polysynthesis means the number of distinct possible word forms is effectively unbounded, so a system trained on limited data meets unseen forms constantly.
This is precisely the situation in which a carefully compiled dictionary, built by people who know the language, outperforms an automatic system — and where the absence of such a dictionary leaves learners with nothing.
Revitalisation
Cheyenne is taught in tribal schools and at Chief Dull Knife College in Montana, with language programmes aimed at moving learners from passive familiarity to active use. The structural challenge is real: a polysynthetic verb system cannot be acquired through vocabulary lists, because the unit of meaning is not the word.
What works is what works everywhere — extended contact between fluent elders and young learners, with the language used rather than described. Documentation supports that; it does not substitute for it.
A dictionary for a language the internet forgot
The free English-Cheyenne Dictionary exists because machine translation does not serve languages like this. Fully offline.
Open the dictionary appFrequently asked questions
What is a polysynthetic language?
One that builds very long words from many meaningful pieces, so that a single word carries what English needs a whole sentence for. The verb typically encodes subject, object, tense, manner and more, and the resulting word is a complete utterance.
How many people speak Cheyenne?
Estimates commonly place fluent speakers in the low hundreds, most of them elderly, across Montana and Oklahoma. It is classified as severely endangered, and revitalisation programmes run in tribal schools and colleges.
Is Cheyenne related to Navajo or Lakota?
No. Cheyenne is Algonquian, related to Cree, Ojibwe, Blackfoot and Arapaho. Navajo is Athabaskan and Lakota is Siouan — three entirely separate families. North American languages are far more diverse than the label Native American suggests.