Authority caveat. Nobody on this site claims to know Navajo authoritatively. Every Navajo form is copied from a named published source with its example number; none is invented. Formal rules are hypotheses to be tested against Grant's texts.

SOURCED stated in a published source we opened HYPOTHESIS our formalization or inference GRANT'S CONJECTURE Grant's own claim OPEN needs data or reading

Theta roles as arguments: the proto-role schema

2026-11-11 audit pass 2. Apostrophe fixes only. Navajo glottal stops typed as ' or ’ (for example in prefix names like ʼa- and ʼeʼe-, or words like Awééʼ) now print as ʼ (U+02BC), and a build step no longer turns the closing quote of an English translation inside a Navajo row into a glottal stop (‘…the girls.’, not ‘…the girls.ʼ). No form, gloss or translation wording changed.

Working note. Kept as written (URL, data and status labels unchanged). Filed under The roles-vs-words thesis; the consolidated argument is the essay roles-not-words. Rule IDs: rules index.

Verbs on this page → dictionary: yiztał ‘kick’ · yishxash ‘bite’

Restated on the unified structure (2026-11-15). The framework is restated in §1: VPISH (arguments merged in the verb’s scope, subject in Spec,vP) with theta roles, and V→v→T with the subject raised to Spec,TP as the working default in tensed clauses; each pattern is what else moves, with an LF whose scope matches c-command. Light d, used below, is withdrawn (Grant, 2026-11-14); light n is an OPEN structural question. Status on the spine: see unified structure, §1. The data, judgments, translations and citations below are unchanged; the analysis below is kept as the record of what was argued then.

What would it mean, formally, for the Navajo verb's argument domain to hold "only a schema of the theta roles … prior to realization as a lexical item"?

On this page: 1 Claim · 2 Definitions · 3 Well-formedness · 4 The yi-/bi- rule · 5 Fourth person · 6 Valence operations · 7 Three theories compared · 8 Strongest version

1. The claim GRANT'S CONJECTURE

"their arguments aren't set within the domain of the VP, but rather only a schema of the theta roles, which can properly be called proto roles here, since they are prior to realization as a lexical item."

Two parts. (a) What occupies the argument positions inside the verb's projection is a role, not a lexical DP. (b) Those roles are proto-roles in Dowty's sense: clusters of entailments, fixed by the event before any lexical item is chosen.

2. Definitions HYPOTHESIS (built from sourced pieces)

Entailment inventory. SOURCED Dowty 1991:572, exx. (27)–(28) (page image checked).

Proto-Agent (27)Proto-Patient (28)
vol: volitional involvement in the event or state
sent: sentience and/or perception
caus: causing an event or change of state in another participant
mov: movement relative to the position of another participant
(indep: exists independently of the event named by the verb)
chg: undergoes change of state
inc: incremental theme
aff: causally affected by another participant
stat: stationary relative to movement of another participant
(dep: does not exist independently of the event, or not at all)

Dowty's Argument Selection Principle (1991:576, (31)): the argument with the most Proto-Agent entailments is lexicalized as subject, the one with the most Proto-Patient entailments as object. Corollary 1, (32), and Nondiscreteness, (34), are on the same page. Dowty adds that the ASP constrains which lexical predicates are possible; it is not a step in a syntactic derivation. That matters below: if the schema is going to do syntactic work, it needs more than the ASP.

Proto-role slot. A slot is a triple

ρ = ⟨ E_ρ , pos_ρ , φ_ρ ⟩
  E_ρ   ⊆ PA ∪ PP         entailment profile, fixed by the event (root + v)
  pos_ρ ∈ {Spec,V ; Comp,V ; Spec,v ; Obl(P)}     first-Merge site
  φ_ρ   = [π: _ , #: _ , rank: _ ]               unvalued at first Merge

Schema. For a root √ and voice/classifier head v, Σ(√, v) = {ρ₁ … ρₙ}, n ≤ 3. First-Merge sites follow Hale's (2003) structural argument positions:

Grant's distinctive restriction.

(G)  No lexical DP is ever first-Merged at pos_ρ for any ρ ∈ Σ.

Exponents. The values of φρ are spelled out by verb-internal pronominal exponents. In the sources these are the conjunct object and subject markers: sh- ‘1s’ and ni- ‘2s’ in Hale 2003 ex. (1); the third-person object forms yi-/bi-; the fourth person ji- (subject) and ho- (object) (Willie 1991 §4.1). They also include null exponents, e.g. the Ø 3rd-person subject in the imperfective (NLA Navajo Verb Modes).

3. Well-formedness conditions HYPOTHESIS

A clause built on Σ is well formed iff conditions C1–C9 hold.

#ConditionStatus / basis
C1Exactly-once valuation. There is a bijection between Σ and the pronominal exponent occurrences in the verb word (null exponents included). Every slot is valued, and no slot is valued twice.HYPOTHESIS The PAH counterpart: argument positions "are satisfied always and only … by clitic pronouns" (Jelinek 1984:44, of Warlpiri). For Navajo, see Hale 2003:29: arguments are "satisfied in situ".
C2Event-fixity. Eρ does not change under the yi-/bi- choice or under fourth-person choice.SOURCED in effect: Hale gives the same translation for the direct and inverse members of a pair ("Conflicting Truths" ex. (2), prior run). The NLA grammar draft, ch. 3 "Basic Clause Structure" (Fernald, prior run), says the two "have the same meaning but have different uses in discourse".
C3Adjunct linking. Each overt DP links to exactly one ρ, and its linear position is fixed by rule set B (topic first), not by posρ.HYPOTHESIS, consistent with W&J 2000 (nominals as topic/focus adjuncts). Refined 2026-10-11 (evening): Willie 1991:51–70 confirms that nouns are adjuncts licensed by pronouns, but position fixes only a default link. The verb's and the noun's meaning can choose another slot (nominal adjunction §11, ADJ-SEM, ENT-LINK).
C4Ranking domain. R(Σ) = {ρ ∈ Σ : π(ρ) = 3}. Voice alternation is defined iff |R(Σ)| = 2 and Σ is transitive.SOURCED W&J 2000:264: the alternation "is excluded where there is a first-, second-, or 'fourth'-person argument, and in intransitive sentences"
C5Agent identification. ρA = the slot in R with the most PA entailments, ρP = the other.HYPOTHESIS (Dowty's ASP applied to the ranked pair)
C6Rank preference. The DP linked to the topic slot should rank at least as high on the status scale as the other DP.SOURCED W&J 2000:273; Creamer 1974 (eight groups); Willie 1991:122 ("must")
C7Inverse definiteness. Under bi-, the DP linked to ρP (the topic) is definite.SOURCED W&J 2000:271, (54)–(56); Willie 1991:126
C8Uniqueness. At most one bi- (inverse) per clause.SOURCED W&J 2000:270
C9Fourth-person exit. If π(ρ) = 4, then ρ ∉ R.SOURCED Willie 1991:122: fourth person is "not involved in the animacy hierarchy"

4. The yi-/bi- linking rule

HYPOTHESIS Given C4–C5, and writing TOP for the slot linked to the clause-initial topic DP:

VOICE(Σ) = yi-   if TOP = ρ_A        (direct)
VOICE(Σ) = bi-   if TOP = ρ_P        (inverse)
subject to  C6: rank(DP_TOP) ≥ rank(DP_other)

Update 2026-10-08, T5: Willie 1991 (39), (72) show that the voice differs, and can be starred, with no nominals at all. So VOICE is restated over the external slot instead of over the topic DP: yi- iff the external slot is ρA, and the external slot is the higher-ranked one. The topic facts follow from linking. The outputs below do not change. See the inverse page, §3.

This is Hale's parsing generalization, restated over slots. In (NP₁) NP₂ X AGROBJ V: if AGROBJ = yi-, NP₂ is the object; if bi-, NP₂ is the subject ("Conflicting Truths" ex. (4), prior run).

Corollary HYPOTHESIS. Let ρrank be the slot whose DP is highest on the status scale. When the two ranks differ, C6 forces TOP = ρrank, so:

bi-  ⟺  ρ_rank ≠ ρ_A

In words: the inverse appears exactly when the referent order (animacy/status, fixed by the lexical DPs) and the proto-agent order (entailments, fixed by the event) pick different slots. When the ranks are equal, both voices are available (Creamer 1974). Qualified 2026-10-14: this holds for animates only. Two equal-ranked inanimates allow no inverse in either order (Witherspoon 1977:66, (11)–(14)); see INV-CAP on the who acts on whom page. Restored 12:46pm: the qualification is withdrawn. In those sentences the moving inanimate outranks the static one (Creamer 1974:36), so they are not equal-rank pairs, and INV-CAP is falsified by Creamer 1974:37 (Eight ranks). The inverse thus diagnoses a mismatch between the schema and its lexical realization. That is about as close as published data come to showing that the slots exist independently of the DPs that later fill them.

Test on the sourced data (forms and glosses as printed):

Ex.FormρA / ρP (by entailments)RankPredictedSource judgment
W&J (61a)łį́į́ʼ tsé yiztałhorse / rockhorse > rockyi-, horse topic✓ ‘The horse kicked the/a rock.’
W&J (61b)tsé łį́į́ʼ biztałhorse / rockhorse > rock✗ (rock topic violates C6)?*
W&J (62a)łį́į́ʼ ʼashkii yiztałhorse / boyboy > horse✗ (horse topic violates C6)?*
W&J (62b)ʼashkii łį́į́ʼ biztałhorse / boyboy > horsebi-, boy topic✓ ‘The boy was kicked by the/a horse.’
Willie (20a)ʼashkii łééʼchąąʼí yishxashboy / dogboy > dogyi-✓ ‘The boy bit the dog.’
Willie (20b)łééʼchąąʼí ʼashkii yishxashdog / boyboy > dog✗*
Willie (20c)ʼashkii łééʼchąąʼí bishxashdog / boyboy > dogbi-✓ ‘The boy was bitten by the dog.’

W&J = Willie & Jelinek 2000:273 (page image checked this run). Willie = Willie 1991:122 (page image checked this run; the typescript renders the nasal hooks as cedillas, and łééʼchąąʼí transcribes its łééʼcha̧a̧ʼí). W&J note: "Speakers vary in respecting the hierarchy, and its use is declining now under language loss" (p. 273). C6 is therefore a preference whose strength varies by speaker. That is not an exceptionless rule, and the ?* marks reflect it.

Where the agent is hosted. SOURCED Willie (1991:123) distinguishes the direct-voice third-person agent, "the external Nominative argument", from "an internal (Ergative) agent, as we see in the Inverse". HYPOTHESIS In slot terms: bi- relinks ρA from the external position to an internal one, and Eρ stays the same (C2). This is a change of pos, not of role. It is the one place where C2 has bite: the voice alternation may move a slot, but it may never rewrite its entailments.

5. The fourth person as an exit from the ranking

Willie 1991 (21)
łééʼchąąʼíshishxash
dog1.3.bit

‘The dog bit me.’

Willie 1991 (22)
łééʼchąąʼíhashxash
dog4.3.bit

‘The dog bit that person.’

Willie 1991:123 (page image checked this run). Compare (20b), * with two third persons.

HYPOTHESIS By C9, valuing ρP as 4 removes it from R, so |R| = 1 and VOICE is undefined (C4). The default (direct) linking then applies even though the agent is lower in animacy. This matches Willie's "the speaker can use a fourth person patient pronoun and avoid having to use an Inverse form" (p. 123; prior-run note).

The schema analysis predicts that exponent choice (π value) feeds the ranking, which feeds voice. So the order of operations is Σ valuation of π, then computation of R, then VOICE. In a derivation, π-valuation must therefore come before the step that fixes yi-/bi-. See D5. Full treatment (Willie §4.2.7, KEEP-DIR vs EDGE, Q–A, coreference, derivations F1–F5): fourth-person page (T8, 2026-10-08 evening).

6. Where the schema could earn its keep: valence operations HYPOTHESIS

If arguments really are slots with entailment profiles, valence operations should be stated over Σ, with no reference to lexical DPs. Hale & Platero (1996) give the sourced contrast:

ADD-CAUSER(Σ):  Σ' = Σ ∪ {ρ_C},  E_ρC ⊇ {caus, vol},  pos_ρC = Spec,v(outer)
  (a) if Σ = {ρ₁} with pos_ρ1 = Spec,V        → spell-out: ł only             (simple)
  (b) if Σ = {ρ₁} with pos_ρ1 = Spec,v        → ρ₁ relinked to Obl;
                                                 spell-out: Qual y + ł + oblique bi-   (complex)
  (c) if |Σ| ≥ 2 (transitive base)            → morphological causative rare;
                                                 analytic (periphrastic) causative     (Kibrik 1993:51 via Gessner)

Update 2026-10-07 (evening), T4: tested against Hale & Platero's lists on the plain ł-causer page. Mov is dropped as a trigger for (b), since motion verbs alternate. The position condition is kept. The entailment claim now sits in a separate linking rule, LINK-1: the sole slot is external iff it is volitional, or sentient with an atelic predicate. That rule fits H&P's whole sample but has not yet been shown to beat a lexical diacritic. Case (c) is relaxed: H&P print a complex causative on a transitive theme.

Update 2026-10-07: a stative base (no chg, no vol) falls outside cases (a)–(c). See MAINTAIN and conditions C10–C11 on the keep vs bring about page.

Case (b) uses Hale 2003's structure for the complex causative: an outer classifier ł, Qualifier y, an embedded vP, and an oblique object (Hale 2003 exx. (17b), (18c), (19); structure read from the prose and trees, Navajo strings not reprinted because the text layer is font-garbled). HYPOTHESIS In entailment terms, (a) applies when the sole slot's profile is P-Patient-dominant (chg), and (b) when it is P-Agent-dominant (vol/mov). OPEN The H&P lists could test whether the entailment profile predicts the split better than a lexical unaccusative/unergative diacritic. For example, are there volitional verbs among the alternating ones? On the English glosses, "roll" and "float" are the cases to check.

7. Three theories compared

(A) Grant's proto-role schema(B) Pronominal Argument Hypothesis (Jelinek 1984; W&J 2000; Hale 2003)(C) Lexical-DP / agreement analysis (Perkins, Platero, Speas, per Fernald 2000 §3.1.1)
What sits in the argument positionρ: an entailment profile with unvalued φa pronominal (D) element, the verb-internal prefixa lexical DP or pro; the prefix is agreement
Overt nominalsadjuncts linked to ρ (C3)adjuncts ("simply optional adjuncts", Jelinek 1984:44, of Warlpiri)arguments that A-move
Inverseρrank ≠ ρA (§4)topic/focus voice over weak pronouns (W&J)object fronting (OSV) plus agreement change
Fourth personπ = 4 exits Rpronoun outside the hierarchy (Willie 1991)as (B), stated over agreement
VPISH movementnone for ρ; DPs are linked, not movednone (in situ; Hale 2003:29)Spec-to-Spec raising of DP
Distinguishing predictionvalence operations and alternations stated over entailments; no D-properties at the θ-position itselfD-properties (definiteness, Condition C behaviour) at the prefixreconstruction of DPs into VP; Condition C effects from VP-internal copies

Honest assessment. HYPOTHESIS (A) and (B) are close to notational variants. The difference is whether what is merged at posρ is a role (A) or a pronoun (B). The sourced facts give the pronominal exponents D-like properties. The fourth person is "exclusively definite" (W&J 2000:278). Under bi-, the topic must be definite (C7). (B) reads this directly. (A) has to say that φ/D features come in with the exponent at valuation, not with the slot. That works, but it adds a step. (A) earns its keep only if some generalization refers to Eρ and cannot be stated over pronouns. The candidates here are the inverse corollary (§4) and ADD-CAUSER (§6). For ADD-CAUSER, see LINK-1 (E) vs (D) on the plain ł-causer page.

8. Strongest formalization reached in v0

HYPOTHESIS The verb's argument domain is a set Σ of slots, each with an event-fixed Dowty profile and an unvalued φ-bundle. Each slot is valued exactly once by a verb-internal exponent (C1), and lexical DPs are linked to slots only as clause-level adjuncts (G, C3). Voice is a function from Σ plus topic choice: bi- appears exactly when the referent-rank maximum and the proto-agent maximum fall on different slots, and only when exactly two third-person slots are ranked (C4–C9). The fourth person takes a slot out of the ranking.