---
title: Whose father? Possessor coreference as an inverse test
date: '2026-10-27'
updated: '2026-10-27'
authors: Grant Golland (research by Grok Bot)
abstract: A possessive 'his' inside one noun phrase can refer back to the other participant
  only in one of the two voices. This looked like the binding evidence the deep-inverse
  question needed. Read first-hand, Aissen 1997 derives the same restriction by ranking
  alone, so the restriction fits any analysis in which the patient outranks the agent.
  This essay assembles the possessor data from Young & Morgan (via Aissen), Platero,
  Willie & Jelinek, Horseherder and Saxon, and shows why they lean toward structure
  without settling it.
status: SOURCED data; D2 (deep vs shallow) OPEN
---

# Whose father? Possessor coreference as an inverse test

*Grant Golland (research by Grok Bot)*, 2026-10-27. Status: SOURCED data; D2 (deep vs shallow) OPEN.

**Abstract.** A possessive 'his' inside one noun phrase can refer back to the other participant only in one of the two voices. This looked like the binding evidence the deep-inverse question needed. Read first-hand, Aissen 1997 derives the same restriction by ranking alone, so the restriction fits any analysis in which the patient outranks the agent. This essay assembles the possessor data from Young & Morgan (via Aissen), Platero, Willie & Jelinek, Horseherder and Saxon, and shows why they lean toward structure without settling it.


*Restated on the unified structure (2026-11-15): the possessor sits in the left specifier of the noun phrase and the possessor prefix is agreement; this essay’s analyses are not yet restated. See <https://navajo-decoded.grant-golland.workers.dev/unified-structure/#us-nominal>. The essay’s data, judgments and citations are unchanged; it is kept as the record of what was argued then.*

<h2 id="2026-11-11-audit-pass-2">2026-11-11 audit pass 2</h2>

Wording only: §3 said “In a VPISH tree”. VPISH is a hypothesis about where the subject is first merged, not a kind of tree, so it now reads “In a vP/TP tree”. The argument is unchanged.

## 1. The question

If *bi-* really raises the patient above the agent (Essay 3), the patient should be able to bind into the agent's phrase. Navajo has no quantifier-binding data for the inverse, but it has possessors. Do they decide the question, labelled D2 on the site?

## 2. The data

Young & Morgan's pair, as cited by Aissen 2000, shows that a "his" inside the acted-on phrase can point to the doer under *yi-* **[SOURCED]**:



> **(1)** *ʼashkii bi-lį́į́ yi-yiiłtsą́.*  
> boy BI-horse YI-saw  
>   
> — Aissen 2000 (27a, b), citing Young & Morgan 1980:171; copied from the working note “Who can "his" point back to under bi-? A possessor test for the deep inverse (T53a)” <https://navajo-decoded.grant-golland.workers.dev/possessor-binding/>, ex. (1a)





> **(2)** **bi-lį́į́ ashkii bi-yiiłtsą́.*  
> BI-horse boy BI-saw  
>   
> — Aissen 2000 (27a, b), citing Young & Morgan 1980:171; copied from the working note “Who can "his" point back to under bi-? A possessor test for the deep inverse (T53a)” <https://navajo-decoded.grant-golland.workers.dev/possessor-binding/>, ex. (1b)





> **(3)** **Bi-lį́į́ ashkii yi-yiiłtsą́.*  
> BI-horse boy YI-saw  
>   
> — Aissen 2000 (29), citing Young & Morgan 1980:171; copied from the working note “Who can "his" point back to under bi-? A possessor test for the deep inverse (T53a)” <https://navajo-decoded.grant-golland.workers.dev/possessor-binding/>, ex. (2)



Platero (1978, page images) gives 'the boy's father saw him', with the boy coreferent, in the *bi-* form. The *yi-* form cannot have that reading **[SOURCED]**:



> **(4)** *Ashkii bizhéʼé yiyiiłtsą́. *  
> boy his:father see   
>   
> — Platero 1978:34, (1.57); 36, (1.62), (1.63); 37, (1.66) (page images checked, 130 and 400 dpi); copied from the working note “What comes first decides: Platero 1978 on who "his" can mean, and a no-movement route to W&J (70b) (T59a)” <https://navajo-decoded.grant-golland.workers.dev/possessor-command/>, ex. (1)





> **(5)** *Ashkii bizhéʼé biiłtsą́. *  
> boy his:father bi:see   
>   
> — Platero 1978:34, (1.57); 36, (1.62), (1.63); 37, (1.66) (page images checked, 130 and 400 dpi); copied from the working note “What comes first decides: Platero 1978 on who "his" can mean, and a no-movement route to W&J (70b) (T59a)” <https://navajo-decoded.grant-golland.workers.dev/possessor-command/>, ex. (2)



Willie & Jelinek's (70) has the owner's name inside the doer's phrase:



> **(6)** *tʼáá ʼałtso Jáan ba-ʼáłchíní dabiiłtsą́*  
> just all John 3POSS-children PL-3ACC-3NOM-saw  
> a. ‘[All John’si children]j saw himk.’   b. * ‘[All John’si children]j saw himi.’  
> — Willie & Jelinek 2000:276–278, (70), (71), (77); copied from the working note “Fourth person: escape from the animacy scale” <https://navajo-decoded.grant-golland.workers.dev/fourth-person/>, ex. (70)



## 3. The argument

A deep inverse predicts this pattern **[SOURCED]** **[HYPOTHESIS]**:

- A "his" inside the doer's phrase can point to the person acted on only under *bi-*.
- A "his" inside the acted-on phrase can point to the doer only under *yi-*.

Platero's coreference rule is precede-and-c-command, with NP and PP as boundaries. It predicts the star on W&J (70b) with no plural condition. In a vP/TP tree, (70b) separates "the patient ends up above and before the doer's phrase" (deep *bi-*, or a null topic) from shallow *bi-* with no topic **[HYPOTHESIS]**.

Saxon's analysis makes the *bi-* landing site A-like by her own criteria. That favours deep *bi-* plus Condition C over a null A′ topic **[HYPOTHESIS]**.

## 4. Counter-evidence

**Aissen 1997.** Read first-hand, Aissen 1997 derives the same owner-outranks-owned gap in Fox, Cree, Ojibwe and Tzotzil by ranking alone. She argues against c-command accounts. The site's earlier verdict POSS-5 ("fits deep") is therefore downgraded to GAP-4: the data fit any analysis in which the patient outranks the agent **[HYPOTHESIS]**.

**The examples use names.** None of the examples uses 'every', so none is a true binding test **[OPEN]**.

**The sources conflict.** Platero (3.18) and Young & Morgan (29) disagree on a coreferent owner inside the doer's phrase under *yi-* **[SOURCED]** **[OPEN]**.

**Horseherder's possessor *yi-*.**
- It needs a third-person antecedent and a different third-person doer that c-commands it.
- It is out when the doer is 'I', 'you' or plural.
- Ranking predicts everything except the plural block, and plural doers are also where W&J (70b) departs from ranking.
- No quantified antecedents are tested **[SOURCED]** **[HYPOTHESIS]**.

## 5. Status

D2 moves from "no data" to "fits deep, not decisive" **[OPEN]**. A new test has been proposed but not found in any source: owner *yi-* on the doer's phrase under *bi-*. Deep predicts ✓ and shallow predicts * **[OPEN]**.


## References

1. Aissen, Judith 1997. On the syntax of obviation. Language 73(4): 705–750. [opened]
2. Aissen, Judith 2000. Full RTF with footnotes (cites Young & Morgan 1980). [opened]
3. Young & Morgan 1980. As cited by Aissen 2000. [not opened]
4. Platero, Paul 1978. Dissertation (page images). [opened]
5. Willie, MaryAnn & Eloise Jelinek 2000. Navajo as a discourse configurational language. [opened]
6. Horseherder, Nicole 1998. A Binding-Theoretic Analysis of Navajo Possessor yi-. M.A. thesis, UBC. DOI 10.14288/1.0088584. <https://doi.org/10.14288/1.0088584> [opened]
7. Saxon, Leslie 2003. WECOL 2001, vol. 13. [opened]

Canonical HTML: https://navajo-decoded.grant-golland.workers.dev/essays/possessors/
