---
title: Where does the word end? One ordering rule, two cycles
date: '2026-10-27'
updated: '2026-10-27'
authors: Grant Golland (research by Grok Bot)
abstract: Navajo puts heads first inside the verb word but last in the clause. Grant
  proposed that the word is the point where the ordering changes. This essay states
  a single ordering rule, CYC, that applies once to the word and again to the clause,
  with the finished word counted as one block. CYC reproduces every sourced order
  tested. It then uses the yes/no enclitic and the question endings to locate the
  sealed unit, which turns out to be the verb word W, not the larger verb-plus-postposition
  complex.
status: GRANT'S CONJECTURE → HYPOTHESIS (CYC); MVS-sealing revised; T44a OPEN
---

# Where does the word end? One ordering rule, two cycles

*Grant Golland (research by Grok Bot)*, 2026-10-27. Status: GRANT'S CONJECTURE → HYPOTHESIS (CYC); MVS-sealing revised; T44a OPEN.

**Abstract.** Navajo puts heads first inside the verb word but last in the clause. Grant proposed that the word is the point where the ordering changes. This essay states a single ordering rule, CYC, that applies once to the word and again to the clause, with the finished word counted as one block. CYC reproduces every sourced order tested. It then uses the yes/no enclitic and the question endings to locate the sealed unit, which turns out to be the verb word W, not the larger verb-plus-postposition complex.


*Restated on the unified structure (2026-11-15): the word edge and prefix order are restated as Grant’s Merge-site hypothesis (GRANT’S HYPOTHESIS, fit/misfit OPEN in §3.2); question enclitics are interpreted in C⁰ and pronounced on a host (H12). See <https://navajo-decoded.grant-golland.workers.dev/unified-structure/#us-word>. 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-1">2026-11-11 audit pass 1</h2>

The earlier argument slid from a model reproducing word order to a demonstrated headedness changeover. That inference is withdrawn: A/B, CYC and the W cycle are hypotheses, not axioms. A sealed pronunciation unit does not uniquely identify a syntactic head or B1/B2/B3. Our primary comparator is V→v→T plus a subject first merged in left Spec,vP and separately raised to Spec,TP; PF/template rivals remain live. The current [boundary and movement tests](/headedness/#2026-11-11-movement-diagnostics) leave the unique boundary OPEN.


<h2 id="2026-11-13-audit-pass-7">2026-11-13 audit pass 7</h2>

The pass-1 note above treats V→v→T and subject raising as a comparator to test against PF/template rivals. Since 2026-11-13 they are the working default in tensed clauses: VPISH frames where the arguments start so that syntax and semantics line up; it is not a word-order account. PF merger and template filling are kept only as labelled contrast, for how exponents line up inside the word, and the open question is what else has to move. The pass-1 note is kept as written. See the [2026-11-13 note](/derivation/#2026-11-13-vto-t-default) and the [pass 7 report](/derivation/#2026-11-13-audit-pass-7).

## 1. The question

In the older model, conjunct prefixes are analyzed as heads ordered first (their headhood remains a hypothesis): Qualifier, then Mode/Subject, then classifier, then stem. Several clause-level items are final; whether each is a selecting head is OPEN: *ya'*, *-go* and *-ę́ę* sit at the right edge. Does Navajo need two rule books? Grant's conjecture is that the word is where the ordering changes **[GRANT'S CONJECTURE]**.

## 2. The data

The yes/no enclitic attaches to a clause-level unit, including a postposition word in front of the verb (NLA):



> **(1)** *Dibéésh nee hólǫ́?*  
> sheep-Q 2-with exist.NI  
> ‘Do you have sheep?’  
> — NLA Navajo Questions, yes/no page (live, 2026-10-21); copied from the working note “Attached or separate? Procliticized vs detached postpositions, where the word ends, and which verb + postposition pairs are idioms (T43e)” <https://navajo-decoded.grant-golland.workers.dev/procliticized-postpositions/>, ex. (4)





> **(2)** *Neésh hólǫ́? *  
> 2-with-Q exist.NI   
> ‘Do you have any?’  
> — NLA Navajo Questions, yes/no page (live, 2026-10-21); copied from the working note “Attached or separate? Procliticized vs detached postpositions, where the word ends, and which verb + postposition pairs are idioms (T43e)” <https://navajo-decoded.grant-golland.workers.dev/procliticized-postpositions/>, ex. (6)



It can sit on any constituent, which then becomes the focus:



> **(3)** *Daʼ shizhéʼé tłʼóół yizhbizhę́ę́sh bee dibé síníloh?*  
> Q 1-father rope 3-3-braid.P-past 3-with sheep 3-2-rope.SP  
> same translation; focus on ‘braided’  
> — NLA Navajo Questions, yes/no page, (9)–(11b), as displayed (brackets are the NLA's); copied from the working note “Can a noun stand between a postposition and its verb? bee dibé síníloh, Willie's uninterruptible verb complex, and where a postpositional phrase is merged (T49)” <https://navajo-decoded.grant-golland.workers.dev/interrupted-verb-complex/>, ex. (10b)





> **(4)** *Daʼ [shizhéʼé tłʼóół yizhbizhę́ę] bee dibéésh síníloh?*  
> Q 1-father rope 3-3-braid.P-past 3-with sheep-Q 3-2-rope.SP  
> same translation; focus on ‘sheep’  
> — NLA Navajo Questions, yes/no page, (9)–(11b), as displayed (brackets are the NLA's); copied from the working note “Can a noun stand between a postposition and its verb? bee dibé síníloh, Willie's uninterruptible verb complex, and where a postpositional phrase is merged (T49)” <https://navajo-decoded.grant-golland.workers.dev/interrupted-verb-complex/>, ex. (11b)



The content-question ending is limited in the inverse:



> **(5)** *Díí=sh háí yiztał *  
> DEM=Q who 3.3.kicked   
> ‘This person kicked who?’  
> — Willie 1991:210–212 (page images); (16), (17), (19), (20) at 300 dpi; copied from the working note “Two question endings compared: content-question =sh and yes/no -ísh. First word or first phrase, and what the inverse forbids (T43f)” <https://navajo-decoded.grant-golland.workers.dev/question-clitics/>, ex. (16)





> **(6)** **Díí=sh háí biztał *  
> DEM=Q who INV.3.3.kicked   
> [‘This person was kicked by who?’]  
> — Willie 1991:210–212 (page images); (16), (17), (19), (20) at 300 dpi; copied from the working note “Two question endings compared: content-question =sh and yes/no -ísh. First word or first phrase, and what the inverse forbids (T43f)” <https://navajo-decoded.grant-golland.workers.dev/question-clitics/>, ex. (17)





> **(7)** *łíí’-ísh iiʼniʼ biisxį́*  
>   
> ‘Did the lightning kill the horse?’ (the only translation printed for (76); no gloss line in the source)  
> — Willie 1991:92 (page image), (76); copied from the working note “Two question endings compared: content-question =sh and yes/no -ísh. First word or first phrase, and what the inverse forbids (T43f)” <https://navajo-decoded.grant-golland.workers.dev/question-clitics/>, ex. (76b)



## 3. The argument

**CYC** **[HYPOTHESIS]**. Inside a unit, the non-heads come first, ordered highest to lowest; then the heads, also highest to lowest. Apply this once to the verb word. Then apply it again to the clause, where the finished word counts as a single non-head.

The earlier CYC comparison claimed to reproduce the following tested strings without movement; its assumptions and exception handling remain hypotheses:

- five Hale 2003 verb words;
- two NLA clause types;
- one Willie & Jelinek sentence.

It also derives Hale's split between lexical heads, which take their argument on the left, and functional heads, which take their complement on the right. That restates Grant’s conjecture within CYC; it does not independently establish the cycle or the headedness changeover.

**The sealed unit is W.** Willie and Hale both say that whether a postposition is attached is a matter of pronunciation **[SOURCED]**. Willie (42) prints *Shił béé hózin*; the NLA writes *shił bééhózin*. The verb-plus-postposition complex (MVS) is a phrase, not a sealed block. It "cannot be interrupted" only because nouns and particles are merged higher. The yes/no ending can still go between a postposition and the verb, as in NLA (6) **[HYPOTHESIS]**.

**The inverse limits shifting, not landing.** (76b) is fine because the yes/no ending is in its plain first position. *(17) is out because the content ending has been shifted off its question word **[HYPOTHESIS]**.

**Two derived results** **[HYPOTHESIS]**:

- The reflexive is two pieces, *'á-* and *di-* (Kibrik). The outer piece sits above *da-*, which gives *'á ⌢ da ⌢ di* by CYC with no new rule.
- Ordinary subjects are pronounced on Mode/Subject by Hale's own portmanteau condition (2003 n. 11).

## 4. Counter-evidence

- **The cycle boundary is stipulated.** Hale says when to stop building onto the skeleton is unsolved (2003 n. 12) **[SOURCED]**.
- **Two exceptions.** Hale flags both himself: oblique objects that jump to the far left (n. 13), and deictic subject prefixes (2001:691) **[SOURCED]**.
- **The interrupted MVS.** NLA (9)–(11b) puts *dibé* between *bee* and the verb, which Willie's rules forbid as written. The site's reading: a PP's site depends on the verb (Essay 8) **[HYPOTHESIS]**.
- **The polar enclitic inside the word.** Hale's *ka-[á]sh nanin[á]* splits two words, not one. The site's "[−MFS]-like" verdict on the prefix zone therefore loses its only datum and goes back to open **[HYPOTHESIS]**.

## 5. Status

Still open **[OPEN]**:

- whether the moved *=sh* lands on the first word or the first phrase (T44a);
- whether any opened form puts *-ísh* inside a verb word (none does, but that is silence);
- the structural questions inside the reflexive analysis R1′.


## References

1. Hale, Ken 2003 (manuscript), pp. 10–15, 27, 43, nn. 11–13. [opened]
2. Hale, Ken 2001, p. 691. [opened]
3. Willie, MaryAnn & Eloise Jelinek 2000, pp. 266–273. [opened]
4. Willie, MaryAnn 1991, pp. 40–50, 64, 92, 210–212, 224 (page images). [opened]
5. Navajo Language Academy. Navajo Questions (yes/no page); Postposition Lexicon. [opened]
6. Kibrik 2005 (reflexive segmentation). [opened]
7. Young & Morgan 1987. [not opened]

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