---
title: 'Roles, not words: what the Navajo verb takes as its arguments'
date: '2026-10-27'
updated: '2026-10-27'
authors: Grant Golland (research by Grok Bot)
abstract: Grant's thesis is that in Navajo the verb's arguments are theta roles, carried
  by the pronominal prefixes, and not the nouns of the sentence. This essay states
  the thesis in three strengths, sets out the sourced Navajo data that bear on it,
  and reports where each version stands. The strong version fails. The middle version,
  sharpened as a feature-bundle hypothesis (FB), meets no contradicting datum in the
  sources opened, but its decisive tests (Condition C minimal pairs, extraction from
  nominals, quantifier scope under the inverse) remain open.
status: HYPOTHESIS (T-STRONG fails; T-MID survives as the feature-bundle version FB,
  with key tests pending)
---

# Roles, not words: what the Navajo verb takes as its arguments

*Grant Golland (research by Grok Bot)*, 2026-10-27. Status: HYPOTHESIS (T-STRONG fails; T-MID survives as the feature-bundle version FB, with key tests pending).

**Abstract.** Grant's thesis is that in Navajo the verb's arguments are theta roles, carried by the pronominal prefixes, and not the nouns of the sentence. This essay states the thesis in three strengths, sets out the sourced Navajo data that bear on it, and reports where each version stands. The strong version fails. The middle version, sharpened as a feature-bundle hypothesis (FB), meets no contradicting datum in the sources opened, but its decisive tests (Condition C minimal pairs, extraction from nominals, quantifier scope under the inverse) remain open.


*Restated on the unified structure (2026-11-15): the “roles not words” theory is dropped; arguments are merged in the verb’s scope with theta roles (VPISH) and the subject raises to Spec,TP by the working default. See <https://navajo-decoded.grant-golland.workers.dev/unified-structure/#us-premises>. The essay’s data, judgments and citations are unchanged; it is kept as the record of what was argued then.*
*Note (2026-11-02): Grant has dropped this thesis. The essay is kept as a historical record; current work is in “Light determiners” <https://navajo-decoded.grant-golland.workers.dev/essays/light-determiners/>.*


## 1. The question

Grant's claim, in his words on the site, is that in Navajo "theta roles, not lexical items, are the arguments". The working notes split it into three versions **[GRANT'S CONJECTURE]**:

- **T-STRONG**: only roles are arguments; nouns play no part in argument structure.
- **T-MID**: the arguments are the role-bearing prefixes; nouns are linked to them, not merged in argument positions.
- **T-WEAK**: roles, not nouns, decide some argument-structure facts (ranking, voice, number).

The question for this essay is narrow: which of these survive the Navajo data that the site has actually opened?

## 2. The data

Navajo verbs are complete clauses without any noun. Willie gives a verb word alone as a sentence, and the same word with an optional noun:



> **(1)** *(ʼashkii) yicha *  
> boy 3s.crying   
> ‘(The boy) he is crying.’  
> — Willie 1991:57–62, (11)–(25); copied from the working note “Which noun is which? Nominal adjunction: nouns attach to role slots, and the verb can overrule word order” <https://navajo-decoded.grant-golland.workers.dev/nominal-adjunction/>, ex. (11)





> **(2)** *yíchxǫʼ  *  
> 3s.ruined    
> ‘It is ruined.’  
> — Willie 1991:57–62, (11)–(25); copied from the working note “Which noun is which? Nominal adjunction: nouns attach to role slots, and the verb can overrule word order” <https://navajo-decoded.grant-golland.workers.dev/nominal-adjunction/>, ex. (12)



A free pronoun can stand with the verb, but not two that both claim one slot:



> **(3)** *shí shiiniltsą́ *  
> I 1s.2s.saw   
> ‘I (am the one) you saw.’  
> — Willie 1991:53–56, (1), (3)–(5), (7)–(10); copied from the working note “Which noun is which? Nominal adjunction: nouns attach to role slots, and the verb can overrule word order” <https://navajo-decoded.grant-golland.workers.dev/nominal-adjunction/>, ex. (9)





> **(4)** **ní shí shiiniltsą́*  
>   
>   
> — Willie 1991:53–56, (1), (3)–(5), (7)–(10); copied from the working note “Which noun is which? Nominal adjunction: nouns attach to role slots, and the verb can overrule word order” <https://navajo-decoded.grant-golland.workers.dev/nominal-adjunction/>, ex. (10)



The distributive prefix *da-* is the sharpest datum. It requires a plural agent in the direct and in the inverse, and in the direct it does not care about the patient (Willie 1991:32, 82–85) **[SOURCED]**:



> **(5)** *ʼashiiké ʼatʼééké dayiiłtsą́ *  
> boys girls PL.3.3.saw   
> ‘The boys saw the girls.’  
> — Willie 1991:81–84, (61)–(66); copied from the working note “Number and quantifiers across the inverse: da- follows the agent” <https://navajo-decoded.grant-golland.workers.dev/number-scope/>, ex. (63b)





> **(6)** *ʼashiiké ʼatʼééd dayiiłtsą́ *  
> boys girl PL.3.3.saw   
> ‘The boys saw the girl.’  
> — Willie 1991:81–84, (61)–(66); copied from the working note “Number and quantifiers across the inverse: da- follows the agent” <https://navajo-decoded.grant-golland.workers.dev/number-scope/>, ex. (63c)





> **(7)** **ʼashkii ʼatʼééké dayiiłtsą́ *  
>   
> (The boy saw the girls.)  
> — Willie 1991:81–84, (61)–(66); copied from the working note “Number and quantifiers across the inverse: da- follows the agent” <https://navajo-decoded.grant-golland.workers.dev/number-scope/>, ex. (63d)



## 3. The argument

Three findings carry the case for T-MID.

First, the prefixes do the work of arguments. Willie (pp. 51–52) and Hale (ms. 5–7, 51) state directly that the pronominal prefixes are the arguments and the nouns are adjuncts linked to them **[SOURCED]**.

Second, *da-* follows the proto-agent role across the voice flip, while definiteness follows the subject position. That is the thesis stated over a real prefix: a role, not a word or a position, is what the prefix counts **[HYPOTHESIS]**.

Third, the ranking that decides between *yi-* and *bi-* is computed before any noun is adjoined: rock and tree swap places with the event, and newborns are re-ranked by construal (see Essay 2) **[HYPOTHESIS]**.

The working notes sharpen T-MID as **FB**, the feature-bundle hypothesis: the verb's argument is a φ-level bundle [person, number, capacity rank κ, role], spelled out as the prefix, and nouns link to it by matching features. FB keeps what works in Jelinek, Willie, Hale and Speas, and gives Creamer's capacity ranking a home as κ **[HYPOTHESIS]**.

## 4. Counter-evidence

T-STRONG fails. Nouns are arguments of postpositions; position and person, not only roles, decide links; quantifier and indefinite facts are hard for it. In that form Grant's claim is wrong **[HYPOTHESIS]**.

FB faces three challenges, none of them a contradicting datum:

1. **Speas.** Speas 1986/1990 places nouns in argument positions. Her Navajo data were checked datum by datum: 0 against FB, 1 supports, 16 compatible, 1 open. Her two tests (no weak crossover; grammatical multiple questions) target an operator-binding version of the Pronominal Argument Hypothesis that FB does not adopt. The 1990 revision and Barss et al. 1989 are not opened, so "survives Speas" is provisional **[OPEN]**.
2. **Condition C.** Hale 1999 (10), (11a) are in tension with FB-5, FB's exemption from Condition C. The tension holds only under ten assumptions that are not established for Navajo. An earlier statement on the site that FB-5 "is wrong" was corrected at 1:07pm on 2026-10-15 as too strong (Essay 6) **[HYPOTHESIS]**.
3. **Possessor *yi-*.** The possessor-*yi-* literature also treats nouns as in argument positions (Essay 4) **[OPEN]**.

## 5. Status

| Version | Status |
|---|---|
| T-STRONG | fails |
| T-MID as FB | survives; no contradicting datum; Condition C (D6) compatible, undecided |
| T-WEAK | survives |

The tests that could still contradict FB are Condition C minimal pairs, extraction out of nominals, and the scope of quantified nouns under the inverse. Willie's quantifier pair (58)–(59) has been checked on the page images, but it does not state the readings that would settle the scope question (T5c) **[OPEN]**.


## References

1. Willie, MaryAnn 1991. Navajo pronouns and obviation (dissertation). Pages cited from page images as recorded on the site. [opened]
2. Willie, MaryAnn & Eloise Jelinek 2000. Navajo as a discourse configurational language. In Fernald & Platero (eds.), The Athabaskan Languages (OUP). [opened]
3. Hale, Ken 1999. Conflicting truths (pp. 171–173 as recorded on the site). [opened]
4. Speas 1986. Dissertation, Navajo chapters (pp. 209–281 as recorded on the site). [opened]
5. Speas 1990. Phrase Structure in Natural Language (revision; Navajo chapter). [not opened]
6. Barss, Hale, Perkins & Speas 1989. [not opened]
7. Creamer 1974. Ranking in Navajo nouns. Navajo Language Review 1(1). [opened]
8. Dowty 1991. Thematic proto-roles and argument selection. Language 67. [opened]

Canonical HTML: https://navajo-decoded.grant-golland.workers.dev/essays/roles-not-words/
