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

Open questions and data needed

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.

Verbs on this page 17 entries in the verb dictionary (show)

yiztał ‘kick’ · néidííłtsʼin ‘hit with the fist’ · yishxash ‘bite’ · yiyiisxį́ ‘kill (one)’ · dabiiłtsą́ ‘see’ · yoolóós ‘lead (one animate)’ · d-ee-zaʼ ‘belch, burp’ · yi-sh-cha ‘cry’ · hólǫ́ ‘exist; (with -í ‘with’) have’ · booyeéł ‘carry (an animate); be carried = ride (in the translations)’ · ní ‘say’ · joolyé ‘be called (named)’ · yiʼash ‘walk (two)’ · nizin ‘think; wish (with an optative clause)’ · hózin ‘be known; with shił béé: I know it’ · yíghááh ‘join; with -tah ‘among’: go among’ · honiigaii ‘the area became hotter; with bitah ‘among it’: he became sick’

Restated on the unified structure (2026-11-15). The open questions are restated in one list on the unified structure. Light d, used below, is withdrawn (Grant, 2026-11-14); light n is an OPEN structural question. The head-direction mismatch argued below as rule sets A/B or a changeover is now stated as Grant’s Merge-site hypothesis and tested in §3.1–3.2. Position-class template statements below describe the order of prefixes; they are not the derivation (§3). Flat-S and other pre-VPISH notation below is labelled contrast. Derivations below in which the verb or subject does not raise are labelled contrast; the working default raises V→v→T and the subject to Spec,TP. Status on the spine: see unified structure, §6. The data, judgments, translations and citations below are unchanged; the analysis below is kept as the record of what was argued then.

What v0 could not settle, and what each item needs: Grant's texts, native-speaker judgments, or a source we could not open.

Does the voice reach the instrument? (2026-10-22, 8:46pm, T50a, T49b–d): every sentence in four NLA Postposition Lexicon lists (instrumental, accompaniment, benefactive, goal; 138 entries, Wayback captures, the live site is down) was checked for third-person subject + third-person postposition object. -ł alternates yił/bił (1992, 294, 2); -á, -aa, -ká, -chʼįʼ, -kééʼ take the y-form; the instrument -ee is bee in both 3/3 sentences (1953; 92, gloss conflict) and no yee occurs. Object-before-instrument order is attested (92, 90). No noun stands between a ranked PP and the verb. With two Ps, the unit P is next to the verb (new IP-8). Status: data SOURCED; P1 HYPOTHESIS, first test passed, weak; IP-7 OPEN. Page: Instrument and voice.

Which postpositions belong to the verb complex (2026-10-22, 12:46pm, T49f, T49h): Willie 1991's grammatical postpositions (28a–i) were checked verb by verb on page images (pp. 41–49, 63–70, 94–96, 100, 208–209). Voice pairs yi-/bi- are printed for ‘for’ with ‘lead’ (85)/(86), ‘to’ with ‘roll’ (83)/(84) and ‘to’ with ‘speak’ (79)/(80), (12a, b), and Willie puts the PPs inside the inverse's scope (pp. 100, 208). The beneficiary and the goal are not participants the stem singles out, so IP-1 is contradicted as worded and replaced by IP-1′ (the voice's ranking decides the site). What makes a role rankable is OPEN: animacy, or relations between participants vs relations to things (IP-7). -ghá and -láah have no examples. CYC is rewritten as v2 (D1 = W, C1′ in the rule) and every derivation gives the same string. Status 2026-10-22, 12:46pm: T49f and T49h done; T49a, T49d sharpened; new T50a–h. See postpositions by verb.

A noun between a postposition and its verb (2026-10-22, 4:46am, T49): NLA (9)–(11b) on the yes/no page print … bee dibé síníloh? ‘Did you rope the sheep with the rope …?’. The object noun stands between bee “3-with” and the verb, although Willie 1991 lists -ee Instrumental as a grammatical postposition (p. 41 (28f), page image) and says the verb complex “cannot be interrupted” (p. 50 (44a)). The site's PC-2 is falsified as written and restated: only a PP whose object is a participant of the verb's role schema must stand next to the verb (IP-1, IP-3). A configurational alternative derives the same string, so the datum is compatible with T-MID but does not decide it. A rule-book check found two more problems: CYC does not order a PP before W once D1 = W (fix C1′), and the preverb rules TP-6 and EN-5 need two preverb sites, which predicts no -ísh on dah after a PP (PV-2). Status 2026-10-22, 4:46am: T45d answered yes for an instrumental bee; new T49a–h. See can a noun stand between a postposition and its verb?

Distance as measure (2026-10-21, 8:46pm, T47): Grant's claim that in motion events the distance phrase (‘a mile’), not the vehicle, is the incremental theme was checked against Dowty 1991 (pp. 567–569, 572, page images), Krifka 1998 (ms.) and Hay, Kennedy & Levin 1999: supported for English (the path is what is used up; the mover and the car are holistic themes; HKL define the incremental theme as the measure along the path and give “?? She ran a mile, but didn't quite finish it”). Navajo: the NLA's entry 479 (‘we ran about two miles …’) is still the only sourced mile sentence; eleven related sentences (how far, as far as, progressive toward a limit with no arrival) and Smith, Perkins & Fernald 2007 on viewpoint and ni- were added; no sentence with both a vehicle and a distance was found. The distance is no slot in the verb word (DM-2, DM-3). Status 2026-10-21, 8:46pm: T47 theory answered; the ‘but didn't get there’ test is designed in schematic form and OPEN; new T48a–h. See distance as measure.

The -ił travel pattern (2026-10-21, 12:46pm, T45f): the NLA Postposition Lexicon's full “transportation” list (17 entries counted, 13 displayed), its “instrumental” list and three random example lists were read in the Internet Archive's copies (the live site did not answer), and Hale 1997 p. 2 and Hale 2003 pp. 14–18 were re-read on preverbs. The pattern is built from the parts across at least six stems and five kinds of carrier (TP-0). The subject is whoever the stem ranks first: the carrier with movement stems, the driver with the ‘drive’ stem -bąs, the traveller with -ee (TP-1 to TP-4). Preverbs dah, yah are the only words seen between a PP and its verb, and they belong to the verb word (TP-6). Status 2026-10-21, 12:46pm: T45f answered as a pattern (strict productivity → T46c); T45d answered for preverbs (rest → T46f); new T46a–h. See the -ił travel pattern.

Attached or separate postpositions (2026-10-21, 4:46am, T43e): Willie 1991 (pp. 24, 25, 48–49, 64, 100, 221, 224) and Hale 1997 (PDF pp. 2, 30–34, 43, 53) were re-read on page images or text layer, and the NLA yes/no page and Postposition Lexicon were opened live. Attached vs separate postpositions is pronunciation over one syntax (Willie p. 64, 224; Hale p. 31). The sealed word-level unit is W, with each PP sealed separately; the MVS is the phrase where they are ordered and the voice is chosen (PC-1), which is why -ísh can sit between a postposition word and the verb. -ísh on an attached postposition is unattested (T45a). Verb + postposition idioms exist with the postposition written apart (NLA bídin nishłį́ ‘I really want (it)’, niyaa hazlį́į́ʼ ‘grew up’, bichʼįʼ sidziil ‘withstand’, -ká + ‘help’), so phrase-level lexicalization does not need attachment (PC-6, PC-7). New row T45a–h.

The two question endings compared (2026-10-20, 8:46pm, T43f): Willie 1991 pp. 92, 210–212 and W&J 2000 pp. 272–273 were re-read on page images, and the NLA content-question page was opened live. On its own host the content-question =sh closes the whole wh-phrase ((19), (20): QC-1), as -ísh does. Willie says “first word” for both endings (p. 92 for -ísh), so the word-vs-phrase contrast in T43f is wording. Only the shifted =sh with a long first phrase is untested (T44a). The inverse restricts both endings, by one rule (QC-5): an ending may sit only on an unmoved host. (76b) łííʼ-ísh and W&J (59) are fine on the inverse subject, while *(17) (shifted =sh) and *(76a) (fronted host) are out. *(76a) and *(13) fall under one domain rule (QC-3). Follow-ups T44a–g: the two question endings compared §9.

Where the yes/no enclitic -ísh can go (2026-10-20, 12:46pm, T41f): Hale 2003 n. 12's two forms were read with a key from the manuscript itself (prose “¡” = á, checked against table (10)'s legible yisdá and the NLA's naniná): ka naash[á], ka-[á]sh nanin[á]. -ísh is a first-constituent enclitic (NLA yes/no), which is Rudin's condition. But Hale's ka stands as its own word, so the split is between words, and the “[−MFS]-like” verdict on the prefix zone is withdrawn (OPEN). New rules EN-0 to EN-6: the host is a spelled-out unit, never a piece of the verb word (a prediction). The neutral position is the first constituent after da’, and head-final order makes (10a)/(10b) one string with two readings. The enclitic's seal is the verb word (B2), not the Maximal Verb Complex (B3). Follow-ups T43a–h: enclitic page §10.

The older definitions behind “edge feature” (2026-10-20, 4:46am, T41e): Chomsky 2000 (“Minimalist Inquiries”) and 2001 (“Derivation by Phase”) were read in full, from OCR of scans checked against page images. Grant's EPP/Edge Feature wording is the 2000/2001 notion. Hiraiwa's Match is MI (40) verbatim. Both texts allow an expletive only at a non-phase head, so the ‘start to’ placeholder moves from the qualifier layer to PINCH (EP-2, restoring IN-8's location). With “Merge preempts Move”, IN-9's two stipulations become theorems (EP-3). The unaccusative default gets MI's defective-intervention mechanism, at the cost of an activating feature on slots (EP-5). MA1′'s reason splits into PIC1 and PIC2 with the same strings. ‘Go’ is not v* by either definition (T41h partly). Follow-ups T42a–h: 2000/2001 page §9.

The borrowed terms checked against their originals (2026-10-19, 8:46pm, T40b): Rudin 1988, Hiraiwa 2001 and Chomsky 2008 were read in full, and Grewendorf 2001 as an abstract. Agree inside the word and Merge at its edges survives. The leftover deictic is now OPEN between Agree and Merge. The ‘start to’ placeholder matches Chomsky's expletive EM (p. 157). The object prefix is not Hiraiwa's Multiple Agree, so SH-6 is favoured. The PP being a phase is our extension, and an inactive-goal alternative is added. Rudin's decisive clitic test points to [−MFS]. U-2 gains a second, theory-internal ground (unaccusative v is not a phase). Follow-ups T41a–h: originals page §9.

Merge and Agree inside the word (2026-10-19, 1:52pm, T40): the word-internal findings sorted by mid-level setting (Grant's typology note). Most inner slots are Agree sites (subject φ on M/S, P's own pronoun, the yi-/bi-/Ø choice); the Merge sites are edges (the bumped object, the leftover deictic, the ‘start to’ placeholder); two filled positions refuse a second occupant ([−MFS]-like); D1 and the PP behave like phases; Multiple Agree is open. Follow-ups T40a–j: Merge and Agree inside the word §10.

‘Start to’ in HMP's own words (2026-10-19, 12:46pm, T38a, T38b, T38d–g): a Hale, Munro & Platero manuscript on the inchoative (hale028, read as text) shows the unaccusative participant's prefix at the far left, before the disjunct ná- ((10a) shináʼniidlóóh), just like ordinary objects ((2b)), because the ’ of ʼnii- fills the object position (“bumping”, n. 2). That answers T38a (the site's ID4 version of eccentric agreement is contradicted; HMP's own version survives) and T38b (the ’ is a IV thematic object). INCH needs a filled object slot “like an ‘incorporated’ postposition”. Verbs with nothing below v (‘cry’, ‘snow’, the seasons) get a role-less indefinite ’. The two unaccusativity diagnostics agree on 5/5 shared verbs. A default non-local Mode (no expletive, no hidden subject) is now favoured. Open: T38c, T38h, T39a–h. Page.

The ‘start to’ test (2026-10-19, 4:46am, T37b): the inchoative prefix ʼnii- leaves the subject of a transitive or unergative verb in place (Hale's handout (10), (11); ʼiʼniishhą́ą́h ‘I start to snore’), but turns the single participant of an unaccusative into an oblique object (yishdlóóh ‘I get very cold’ → shiʼniidlóóh ‘I start to get very cold’; Hale PA paper (51), handout (12), page images). So the low participant's subjecthood is derived and can be bled, and the prefix sorts ‘sleep’, ‘snore’ and ‘get cold’ as LINK-1 does. U-2 (Mode takes the closest eligible slot) covers both forms with no movement and lets the prefix's own indefinite ’ supply the third person Hale finds in Mode, so no expletive is needed; the handout's raising-plus-eccentric-agreement is costlier but not contradicted. Open: an unaccusative ‘start to’ form with a disjunct prefix (T38a); subject or object ’ (T38b). Page.

Subject on the head (2026-10-18, 8:46pm, T35c, T36g): ordinary subject prefixes are pronounced inside the Mode/Subject head because Hale 2003 n. 11's portmanteau condition picks out exactly the slot VPISH puts in Spec,v; that derives WF-6. The fourth-person, indefinite and areal subjects (Hale 2001 (22a–c), read on page images) precede the qualifier, so they are not M/S exponents. They are the deictic part of the subject slot, left over when Mode takes person and number, and adjoined to the highest head (as Hale 1997 and 2001 placed them). This predicts that only subjects have a separate deictic position (4th-person objects ho-/ha- do not) and explains why a 3rd-person object has no prefix under ji- (Willie (23b) jishxash) but yi- under a noun subject (Willie p. 129 ʼashkii yishxash). Open: is φ under ji- really 3rd person (T37a); does a one-place labile verb's argument move up inside the word (T37b); the indefinite subject with a 3rd-person object (T37c). Page.

Reflexive vs reciprocal (2026-10-18, 12:46pm, T35b): the reflexive is two pieces, an outer ʼá- above plural da- and an inner head di- (Kibrik 2005, from Young & Morgan 1987:70–71; Young & Morgan's template via McDonough 2019). The reciprocal ʼahi- is one piece in the object position, below da-. CYC derives both orders, and the WD6 clash is gone. The inner di- is shared with the agentive passive. Open: Young & Morgan's own position for ʼahi- (T36a), di--only reflexives (T36b), why the reciprocal lacks di- (T36c), whether the outer piece is the object itself or a binder (T36d, T36f), scope against da- (T36e), WF-6 after the passive split (T36g), reflexive over causative (T36h) (reflexive vs reciprocal §11).

Willie's framework (2026-10-18, 4:46am, T27o): Willie 1991 Ch. 1–2 read on page images. There are no trees, and the framework is stated outright: Jelinek's pronominal-argument proposal in Government-and-Binding terms (pronouns in A positions leave traces, and the chain gets the theta role and Case; nouns are in A′ positions). Restated, her “pronoun (chain)” is the site's role slot, and her three consequences (no Case on nouns, adjoined clauses, no movement to or from argument positions) follow from nouns entering only after the verb word's cycle. Open: the yi- of ‘I am grinding it’ and ‘I saw it’ (T35a); why the reflexive precedes da- but the reciprocal follows it (T35b); and more (Willie's framework §10).

Relative as CP (2026-10-17, 8:46pm, T27m): Platero 1973's island data (Ch. II, page images) show the relative tie behaves like movement (it obeys coordination and relative-in-relative islands; coreference does not), but the thing that moves is a silent copy of a pronoun slot, never the noun. Only slots can be relativized (subject, object, postposition object, possessor; not the bare noun of an enclitic phrase). Raising of the slot is adopted; matching survives only if the enclitic is a complementizer. Open: reconstruction data (T34a), D or C (T34b), the possessor vs (104) (T34c), the speaker split on nominalized clauses (T34d), leaving the MVS cycle (T34e), slots vs full noun phrases (T34f), and more (relative as CP §10).

One-place verbs + postposition (2026-10-17, 12:46pm, T32e): Willie p. 97 (“no grammatical PPs” in intransitives) is reconciled with p. 94 by reading it for the whole MVS. The real question is which one-participant verbs let the postposition's object join the frame. Willie's animacy condition fails on (84); an extension of Dowty (LIC-PROTO) and an animate-external-argument condition (LIC-EXT) both fit. (93) shows two grammatical PPs, and (92)/(93) put a noun inside what should be the MVS. Follow-ups T33a–h: one-place verb + P §11.

Postposition units (2026-10-17, 7:42am, T32): what category the postposition is (sources: lexical P, head-final, with an agreement-bearing pronoun; nobody opened calls it T or C; light p not found in any Navajo source), verb + postposition as one unit, the inverse bi- on P, and phrase-level lexicalization (Willie (35); Chee & Yazzie’s “phrasal verbs”); follow-ups T32a–j: postposition units §9.

Clause spine (2026-10-17, 4:46am, T27l): the MVS, topic/focus nouns, fronted question words and the =go clause placed in a Qual > M/S > v > V spine with silent Foc/Top above it; W&J 2000 name no such heads; “parallel processing” split into grammar (cast identity, focus-to-focus mapping) and parsing (Platero 1973, opened); W&J (49) prints R2 indices for Willie's (9a); follow-ups T31a–h: clause spine §11.

The inverse as Agree and EPP (2026-10-16, 8:46pm, T27k): yi-/bi- restated as probes and an EPP feature (Hale 1997, newly used; Bruening 2005), with four options derived step by step; raising (inverse only) and in situ both survive, inverse linking is disfavoured by da-; follow-ups T30a–h: the inverse as Agree and EPP §10.

Rule sets A and B as spell-out (2026-10-16, 12:46pm, T27j): A and B restated over Hale 2003's hierarchy as one ordering rule applied twice, first to the word and then to the clause (CYC), and tested against Kayne's antisymmetry and Goad & Travis's prefixation; follow-ups T29a–j: A/B as spell-out §11.

Condition C by site (2026-10-16, 4:46am, T27i): Hale's (10), (11a,b) and Speas's (4.119) re-derived at CP, QualP, M/SP, vP, VP, PP and left-edge sites with derivational c-command; no site predicts both; follow-ups T28a–f: Condition C by site.

Framework audit (2026-10-15, 9:50pm): which sources and which of our own rules are pre-VPISH, transitional or minimalist, with restatement tasks T27a–p: framework audit.

Flagship test list: the open items TA-a to TA-h for the central claim are on the theta-args thesis page.

1. Decisive tests for the two formalizations

Update 2026-11-04 ~4:50am PT: on pos-labels: (b) answered (HYPOTHESIS; data SOURCED): the "Pos. IV" of ʼahééʼa- 49/14 is a label of the prefix entry. The same morphemes are "Pos. IV-V" at 44/17, both entries head causatives and intransitives, and 42 of 46 passenger verbs cite IV-V/IV/V prefixes. So the label is no evidence of object position, and T87a stays open. (d) in part: 45/2 shows the same field/parenthesis mismatch with plainly Ø forms. (a), (c), (e) still open.
#QuestionWhich rule it testsWhat would settle it
T1Are the clause-level particles da’, ya’, -go, -ę́ę, -sh, lá, -shąʼ and the doo … da frame heads or phrasal operators/enclitics?(i′) vs Hale's extended-projection head-initiality (headedness (i), D6)Texts. Embedded questions with da’? Co-occurrence and ordering of several final particles? Selection effects (does the particle determine the clause type of its complement)? 2026-10-10 (afternoon): for the relativizer =yę́ę/=ę́ę, Willie 1991:153, 158 argues it is the head of the relative (“a demonstrative pronoun”), binding a slot inside the verb (relatives §11). 2026-10-15 (4:46am): Hale & Platero 1974:27 likewise treat -ę́ę/-ígíí as a determiner, and (51) shows a word-internal COMP -í- followed by phrasal enclitics in one phonological word. Both favour (i′) for this particle (H&P 1974 §6).
T2Where exactly is the head-direction switch: B1 (# disjunct/conjunct), B2 (verb word / skeleton domain) or B3 (Maximal Verb Complex)?Grant's ∃!B claim (headedness §3)Find any functional head between B1 and B2 or between B2 and B3, and read off its direction. Datum 2026-10-11 (afternoon): between B2 and B3 the order is head-final (PPs before the verb word, object pronoun before the postposition), yet the stretch belongs to the verb's role schema (Willie 1991:50, (44c)). That suggests two boundaries, ordering at B2 and schema at B3 (HYPOTHESIS WORD-EDGE′, pleonastic & psych §9). Datum 2026-10-11 (evening): the default links between nouns and argument prefixes are nested around the MVS edge (B3), with the postposition's object counted as the innermost slot (Willie 1991:59–68, (17), (20), (31); HYPOTHESIS MIRROR-LINK), and nouns adjoin to the MVS as a whole (MVS-UNIT). Three properties now point at B3, one (prefix order) at B2 (nominal adjunction §8, §10). Datum 2026-10-18 (12:46pm): the reflexive is split across B1. Its ʼá- is disjunct (Young & Morgan Ic) and its di- is conjunct (Kibrik position 8), and binding needs both and the subject in one domain. So B1 cannot be a cycle boundary (HYPOTHESIS, reflexive vs reciprocal §10). Datum 2026-10-19 (4:46am): the inchoative's oblique object is a null-postposition phrase that is “sometimes … conjunct …, sometimes … disjunct” (Hale 2001:683 n. 4); in handout (23a) it precedes the disjuncts náá- and da-. Another element that straddles B1 (HYPOTHESIS, the ‘start to’ test §10).
T3Inside the word, does a Qualifier meaning scope over the mode/aspect meaning or under it? Does Obj scope over Subj or the reverse?Hale (Qual > M/S, head-initial) vs Rice (scope to the right, head-final); V-P object raising (D2)Native judgments on scope-ambiguous verb forms. This is Grant's open question 6 below.
T4Does the simple ł-alternation track the entailment profile of the sole argument (chg vs vol/mov), or a lexical unaccusative/unergative diacritic?ADD-CAUSER (schema §6)Check volitional-looking verbs (e.g. the "roll", "float" glosses) against the H&P lists. Look for alternating verbs whose sole argument is volitional. Partly answered 2026-10-07 (evening): no single entailment works; a two-entailment rule (sentient + atelic, plus volitional) fits the whole sample but has not yet been shown to beat a diacritic. See the plain ł-causer page and its new questions T4a–T4f. Datum 2026-10-19 (4:46am): a second, independent diagnostic, the inchoative ʼnii-, sorts ‘sleep’, ‘snore’ (subject kept: starts in Spec,v) and ‘get cold’ (participant made oblique: starts in Spec,V) exactly as LINK-1 predicts (HYPOTHESIS IN-6, the ‘start to’ test §5; T38d runs it over the whole sample).
T5Does the inverse change the agent's position (Willie's "internal (Ergative) agent") or only linking (W&J's constant 3ACC-3NOM gloss)?V-P movement in D4Binding or reconstruction asymmetries between (20a) and (20c)-type pairs. Partly answered 2026-10-08: Willie 1991 §3.3 (pp. 71–100) and Sandoval & Jelinek 1989 both say position. The inverse makes the agent the internal argument and the patient (or goal/benefactive) the external one. The choice is made inside the word, before any nominal is linked: bare yiztał/biztał differ in meaning, and Willie's (72) stars need no nominals. Whether VPISH raising is real here still needs scope/binding evidence. See the inverse page and its questions T5a–T5g.
T6Is the rank preference C6 categorical for some speakers and gradient for others?C6 (schema §3)W&J note variation and decline. Texts by speaker and age, with an inverse count per Creamer group pair.
T7Is there any generalization stated over entailments (Eρ) that cannot be restated over pronouns?Grant's schema vs PAH (schema §7)If none turns up, (A) and (B) are notational variants.
T8Is the fourth person (ji-/ha-/ho-) another ranking-driven inverse, or an escape from the animacy scale?C9 / EXIT; KEEP-DIR vs EDGE (schema §5, D5)Answered 2026-10-08 evening: Willie 1991 §4.2.7 and W&J 2000 §11.4.6 — escape hatch, not a second inverse. Fourth is off the scale; marking the human patient fourth keeps the agent as external subject (direct), instead of flipping to bi-. See the fourth-person page and T8a–T8g.
T10What does number marking inside the verb word read across the inverse: the subject position (EXT) or the agent role (ρA)? And does a quantifier on the inverse subject scope over the internal agent (T5c)?DIST-1–5, Q-1 (number & scope §3); V-P vs V-H (§5)Partly answered 2026-10-09: Willie 1991:32, 78–85. da- counts the agent in both voices, while definiteness reads the subject position. So the word carries both role-indexed and position-indexed requirements. The scope pair (58)–(59) is page-checked but not decisive. See T10a–T10f on the number & scope page.
T12Across a =go “when/if” clause and the clause after it, what must stay the same: the referents, the roles, or the positions? And how does the fourth person mark a new participant?GO-INTR, GO-PP, GO-SWAP, GO-NO-NEW, GO-4-AGT, GO-PAT, GO-4-BOTH (=go chains §4)Partly answered 2026-10-09 (evening): Willie 1991:132–140. In transitive pairs the two referents are fixed and only their roles can change. π = 4 can bring in a new agent only from the first clause, and the patient must stay unless the new one is indefinite. Her printed indices in (9) conflict with her prose (T12a). See T12a–T12f. Update 2026-10-10: T12f answered for reported speech (with all third persons the subjects must differ, so GO-PP is specific to =go); T12a tilted toward R1 by Willie p. 149 (“coindexed according to theta role”), still open; T12c: Willie's reason found (a second-clause 4th person needs a first-clause antecedent, p. 149). See the reported speech page §8. Update 2026-10-10 (afternoon): T12f answered for relatives: yes, parallel processing holds (Willie 1991:166, (27)–(31)), and the main verb's yi-/bi- decides which role the relative fills. T12a: in relatives, 4/4 coreference follows the person value across a patient/agent split ((49), (52)). See the relatives page §11.
T13In “X said that Y …” and in person-focus sentences, what does a person value on a role slot decide: who the referent is, and what kind of thing (person or event) can fill the slot?RS-1, RS-33, RS-44, RS-Q, RS-RESP, RS-4×1, RS-4×3, RS-2, FOCUS-4; PI-MATCH (reported speech §5–6)Partly answered 2026-10-10: Willie 1991:114, 140–150. All-3rd under ‘say’ forces different subjects; 4/4 gives the same subject; 3 never picks up 4 in either construction. A 4th-person object of ‘see’ can only be a person, never an event. Open: respect vs quotative as one morpheme (T13a); the extra 2nd-person slot of nijiní (T13b); booyeéł and the inverse (T13c); other constructions (T13d–f). Update 2026-10-10 (afternoon): T13d answered: the 3/3 default is set by the construction, with three values: same (=go, relatives), different (‘say’), either (‘believe’ + =nígíí). One 4 = different and 4/4 = same in all four (relatives §9). Update 2026-10-10 (evening): T13a narrowed: Willie treats respect, reportative, story, impersonal and optative as uses of one fourth person (pp. 113, 121–122), told apart by distribution, not form. PI-MATCH (i) fails under the optative: in (18) a 4 subject under ‘think’ is the 3rd-person thinker (fourth person uses §11).
T14In relative and nominalized clauses, what picks the argument the relative names (‘the one who …’), and why does the 3/3 default differ between adjunct and argument clauses?REL-HEAD, REL-ABS, REL-EXT, REL-NR, REL-4-HEAD, REL-EDGE, REL-ADJ, REL-PP, REL-NOUN, REL-4, BEL-3/3, BEL-3/4, BEL-4/4 (relatives §8)Partly answered 2026-10-10 (afternoon): Willie 1991:151–181. The relative names the internal argument (patient in the direct, agent in the inverse), the external one when both arguments have nouns, and never a 4 in a restrictive relative. Open: T14a–T14f. Update 2026-10-10 (evening): T14a gains a third complement-taking verb: under ‘think’ (nizin) with an optative clause, the clause's 4 subject is the thinker (Willie (18)–(19)), unlike ‘say’ and ‘believe’ (T15e). Update 2026-10-15 (4:46am): T14e partly answered from Hale & Platero 1974. In the printed data the post-relative-head claim is a difference of analysis, not data, and the real speaker split concerns head choice when both nouns are inside the relative (page; T23a–b).
T15What is the fourth person for, and when may a noun be attached to it?NOM-4, PRO-MATCH, X-MOTHER, X-REPORT, X-STORY, RESP-JINI, IMP-4, OPT-4, OPT-PROH, ONE-4; OPT-LINK (fourth person uses §9–11)Partly answered 2026-10-10 (evening): Willie 1991:109–122. Nouns count as third person and cannot attach to a 4, except for respect to one's mother (8), the reportative with léiʼ (11) and story openings (14). The five uses share one job: avoiding naming or pointing at anyone. The optative 4 takes the referent of the main verb's thinker, whatever its person. Open: T15a–T15g.
T16When a noun is put inside the verb (possessor raising), does it take an argument slot? And where exactly is the word edge between rule set A and rule set B?PR-OBJ, PR-NOARG, PR-TRANS, PR-INAL, PR-VAR, PR-AMB, POSS-YB, POSS-DA, LEX-PP, PR-ANIM, PR-3ONLY; PR-SLOT, WORD-EDGE (possessor raising §9)Partly answered 2026-10-11 (early morning): Willie 1991:182–198. The incorporated body part is “not an argument”; the verb keeps two slots and the possessor fills the object slot. One string can be one verb word or noun + verb, with different meanings and no phonological difference, so the A→B edge is structural (HYPOTHESIS WORD-EDGE). Open: T16a–T16h, incl. whether Willie's animacy account lets a non-argument count for ranking (T16c).
T17Where does a psych verb put the one who feels, and is a “pleonastic” subject slot a role? Does the verb's role schema end at the verb word or at the verb-plus-postpositions unit?HO-PLEO, HO-NONOUN, HO-LOC, PSY-EXP, PSY-HO, PSY-KNOW, IND-NONOUN, AG-PASS, P-LEFT, IMP-A, GPP/LPP, MVS; SLOT-TYPE-LINK, EXP-OBL, WORD-EDGE′, ROLE-ORDER (pleonastic & psych §10)Partly answered 2026-10-11 (afternoon): Willie 1991:35–51, 102–108; Dowty 1991:579–580. The Experiencer is always the object of the postposition -ł ‘with’, never the subject; Dowty's tie account allows that. In ‘know’ both participants are on postpositions and the verb word has only ho-. What may be adjoined depends on the slot's value type (3 → noun, areal ho- → place phrase, indefinite → nothing). WORD-EDGE revised to WORD-EDGE′. Open: T17a–T17i, incl. whether ho- is a slot with no role, which would cut against Grant's thesis (T17a).
T18What decides which noun goes with which role slot: position, the slot's entailments, or knowledge about the world? And around which boundary are nouns ordered?ADJ-NOTARG, ADJ-LEFT, ADJ-ONE, ADJ-PHI, ADJ-FREE, ADJ-OBJ, ADJ-TWO, ADJ-PP, ADJ-SEM, ADJ-SUBJ, ADJ-AMB; SCHEMA-PARSE, MIRROR-LINK, ENT-LINK, MVS-UNIT (nominal adjunction §11)Partly answered 2026-10-11 (evening): Willie 1991:51–70. Nouns are never arguments, only adjuncts licensed by a pronoun. Position gives a default link (lone noun → object; innermost → PP object; outer → subject). The verb's and the noun's meaning can override it. The number of slots sets how a noun string is parsed. Overrides mix slot properties (round/plural objects) and world knowledge (“horses typically are led”). Open: T18a–T18h, incl. the unexplained (25) ‘shooting’ (T18a) and *(33c) vs (28) (T18b).
T19When you ask “who?” in Navajo, do you ask about a word or a role slot? When can a lower argument's question word go ahead of the agent, and which orders are impossible?WH-ADJ … WH-LOCAL; WH-SLOT, FOCUS-RECOVER, MIRROR-LINK′, WH-SPLIT, LOCAL-Σ (wh-questions §11)Partly answered 2026-10-13: Willie 1991 Ch. 8, pp. 199–218. A question word is an adjunct bound to a 3rd-person slot, never to 1st, 2nd or 4th, and it gets its slot by the same order and voice rules as nouns. It can go ahead of the agent only when the roles cannot be confused (‘What did John see?’); with two humans, háísh John yiyiiłtsą́ is only ‘Who saw John?’. That conflicts with W&J 2000 (58b) (T19a). No adverbial and argument question word in one clause *(27). Open: T19a–T19h.
T20What decides who can act upon whom: weighted proto-agent potential, the event's entailments, or the noun? And can a participant with no capacity to resist ever be the inverse subject?INV-3′, INV-CAP, RANK-W, RANK-EV (who acts on whom §8)Speaker judgments on bi- for equal-ranked animates; a search for an inanimate inverse subject; Hale 1973 itself. T20a–T20f on the page. T20b answered 12:46pm: yes, Creamer 1974:37 Tʼiis są́ biisxį́; INV-CAP withdrawn (Eight ranks).
T21What exactly sets Creamer's eight ranks: potential proto-agent properties, the event's entailments, or relatedness to people? Why does old age rank below a tree, and does lightning outrank animals only with ‘strike’?RANK-G, RANK-EV′, RANK-LT, RANK-INF, RANK-PERS (Eight ranks §7)Two non-moving inanimates; group 8 nouns with other verbs; lightning with ‘kill’ vs ‘strike’ by the same speakers; Young & Morgan on -kʼi, -nahjįʼ, -ił. T21a–T21e on the page.
T22Do Speas's configurational data (1986) ever require a noun in an argument position? Is Condition C ever the only thing that excludes a reading, and are Platero's positional generalizations grammar or parsing?FB-3a LINK-NEAR, FB-6 PAR, FB-4, FB-5 (Speas 1986 §6)Equal-rank Condition C pairs; weak crossover with háí; Speas 1990 Ch. 4; Platero originals. T22a–g on the Speas page. 2026-10-15: T22c partly answered. Hale & Platero 1974 (44) is a parsing principle applied “with varying degrees of strictness”, so LINK-NEAR(a) is a default (FB-3a(a′)) (page). 2026-10-15 (12:46pm, corrected 1:07pm): T22b partly answered by Hale 1999 (10), (11a,b): readings with two people are excluded, which is in tension with FB-5 conditional on assumptions A1–A10; Speas's (4.119) points the other way (page, audit). OPEN
T23Hale & Platero 1974: does a noun after the relativizer ever head the relative where Willie's parse gives a different meaning? What do the unprinted island data show? Is -ę́ę/-ígíí a determiner?FB-3a(a′), FB-7 SURF-LINK, REL-D, B-ENC (H&P 1974 §8)T23a–g on the H&P 1974 page: the decisive (12)-type judgment, the IJAL 1974 article, REL–N vs N–REL order, island data, lenient (32), COMP -í-, spellings. OPEN
T24Hale 1999: Condition C holds in Navajo except under Platero's parallel parse. Why must the inner verb be transitive? Why do Hale (11a) and Speas (4.119) disagree? Is the disjointness Condition C or a fourth-person effect? Is yi- a subject prefix? With nouns inside the clause, what still separates FB from Speas?FB-3c LINK-IN, FB-5′, FB-6′ (Hale 1999 §6)T24a–f on the Hale 1999 page and T25a–h on the derivation page. T25c not yet answered (8:46pm partial, revised 9:20pm): Jelinek 1984 tree (28) is pre-VPISH notation, no verdict on A3 (T26g); Baker later n. 16 contemplates PP/vP (against A3); Baker 1996 unopened (page). Still open: judgments, Platero 1978/1982, Hale & Platero 1996, Uyechi 1996, D8. OPEN
T26Attachment height for Navajo nouns: Baker 1996 trees; Willie & Jelinek 2000 topic/focus height; postpositional objects as low adjuncts vs P-arguments; c-command from a prefix; restating the PAH in VPISH/TP/CP structure (T26g); m-command reading of the literal Jelinek (28) diagramA3, FB-5 (Jelinek & Baker attach §11)T26a–g on that page. T25c is not yet answered (revised 9:20pm: Jelinek's tree (28) is pre-VPISH notation; no verdict). OPEN
T26g(new 2026-10-15, 9:20pm) Restate the pronominal-argument hypothesis in a VPISH/TP/CP structure and re-ask where nouns attach (above Spec,vP, at vP, at PP, or in a left-peripheral topic/focus position). Check what Baker 1996 and later work do (the polysynthesis-parameter structure) and whether a minimalist reanalysis of Jelinek exists, citing only what is openedA3, FB-5, Hale 1999 target (Jelinek & Baker attach §12)Baker 1996 (not opened). Progress (9:50pm correction: W&J 2000 pp. 267–269 were opened; trees (43)–(48) put nouns as S-adjuncts above IP; Baker's Kinande trees (16), (19) pp. 14–17 put dislocated nouns at TP/CP): Hale 2003 ms. (opened in full) gives a Spec,v/Spec,V restatement with nouns “adjuncts to the clause” (p. 51) but no tree for nouns; Ormston 1993 and Frantz's review give Baker secondhand. OPEN
T30a–h(new 2026-10-16, 8:46pm) After restating the inverse as Agree + EPP: is the goal/benefactive PP above the theme (decides whether INV-5 follows from closeness)? Does Mode/Subject agree with the agent in the inverse? How does da- find the agent (role probe, bi- as agreement head, or concord through M/S)? Scope in (58) (raising vs in situ; = T10a). Does the direct have EPP raising? Hale 1973 itself on topic. Landing site Spec,M/SP vs Spec,QualP. Is the object prefix a pronoun or an agreement head?Three-participant verbs; Young & Morgan 1987 paradigms; speaker scope judgments; Hale 1973; Hale & Platero 1996; Speas 1990 ch. 4 (details)
T31a–h(new 2026-10-17, 4:46am) After placing the clause in a spine: is the =go clause adjoined to matrix FocP or TopP (any sentence with a matrix topic noun and a =go clause)? R1 or R2 for yiztał=go nábidííłtsʼin (Willie (9a) vs W&J (49); = T12a)? Is the lone 3rd-person slot beside a fourth person topic or focus? Is linking to a deep [foc] slot Agree or interpretive? How does IA-1* spell bi- on a postposition ((50b))? Do lexical PPs belong to the MVS? Platero 1976 and the 1974 IJAL printing. Why does the inverse block wh fronting ((60b))?Grant's texts; speaker judgments; Willie Ch. 2–5; Platero 1976 (details)
T32a–j(new 2026-10-17, 7:42am) After collecting what the sources say about the postposition: does any Navajo source place P under a light p or call it functional (Speas 1990, Hale & Keyser 1993, Hale & Platero 1996)? PU-L or PU-G (two grammatical PPs in one clause; -tah/-ta’ as both kinds)? Is =go related to =góó? Does “theme in Spec,PP” extend beyond drift-away verbs? Willie p. 97 vs p. 94 on intransitives; are the Chee & Yazzie verb + postposition pairs stored, and are they in Young & Morgan? Speas on b-; IA-1* spell-out of bi- on P; nouns between a bare-noun P and the verb; testing “phrasisized”Postposition units §9
T33a–h(new 2026-10-17, 12:46pm) After T32e: LIC-EXT vs LIC-PROTO (an inanimate mover with an inanimate P object: y- or not?); is Witherspoon's ‘upon’ a postposition (T20c)? Willie (28a) -aa “only with animates” vs (83); which PPs are inside the MVS in (92)/(93)? Is haasʼnaʼ a transitive stem? *yitah honiigaii? Does the comitative count in the dual? Two grammatical PPs under PU-GOne-place verb + P §11 Update 2026-10-31 (~9:30pm PT): T33e answered on indef-subject: Young & Morgan list the climbing verb as haashʼnééh, d-classifier, an intransitive with a postpositional object ("Bąąh haashʼnééh, to climb it (a tree)", p. 434 def 8; SOURCED), so haasʼnaʼ is not a transitive stem (TI-12).
T34a–i(new 2026-10-17, 8:46pm) After T27m: raising vs matching by reconstruction with a pronounced outside head (idiom chunk, bound pronoun)? Is -ę́ę/-ígíí D or C in relatives? Is the island only a CP under D (Platero (104) vs his possessor relatives (51)–(53))? Why do speakers split on relativizing into an -ígíí clause (?(96), ?(100))? How does the slot copy leave the MVS cycle (joins T31d)? Slots vs “full noun phrases” as what can be relativized (separates the adjunct, inside-clause and STRATA views)? Is H&P's -í- a slot; Kayne ch. 8? Does Platero 1974 keep the 1973 judgments? Is an adjunct =go clause an island?Relative as CP §10; speakers; Platero 1974; Young & Morgan; Kaufman 1972
T39a–h(new 2026-10-19, 12:46pm) After T38: page images of the HMP manuscript (restored stems, tree layouts); any sign that the unaccusative ‘start to’ form has a third-person subject (A4 vs A5); which Position I prefixes block the leftward shift, and why the far left; the ʼa- of ʼahoʼnii-, indefinite object or ‘away’; weather and season verbs, zero-place or not; can an incorporated noun satisfy INCH (STRATA says no); the 2000 printed version; why momentaneous stems only.‘Start to’ in HMP's words §11; a reachable copy of hale028; Y&M 1980G:320–321; Hale, Munro & Platero 2000; Smith 1996; texts
T90a–e(new 2026-11-03, ~8:55pm PT) From driver-3i: (a) does a noun-less Ø driver verb (45/8) differ from its 3i twin (49/15) as definite pro vs indefinite 'something'?; (b) is the intransitive vehicle 3i in object position (Y&M label ʼahééʼa- "Pos. IV, 3i obj." on intransitives 50/1, 50/2, vs adii- 28/6 "Pos. V"; cf. T87a)?; (c) what is the 3i row of the p. 50 table?; (d) is 45/1's prefix ʼahénáʼa- or ʼahéná-?; (e) does a 3i driver verb ever take a bee-phrase naming the vehicle?Y&M grammar on Pos. IV vs V; print pp. 45 and 50; wider bee/yee sentence searches. OPEN
T91a–e(new 2026-11-04, ~4:50am PT) From pos-labels: (a) do 49/14 (Pos. IV) and 44/17 (Pos. IV-V) differ in print?; (b) by what rule do Y&M assign IV, IV-V or V to a 3i prefix entry?; (c) is there a source-backed Navajo unaccusativity diagnostic that can be run on passenger 3i verbs (49/19, 50/2, 286/19)?; (d) which homonymous entry do 10 passenger verbs cite (ʼadii-((1)), ʼani-, ʼa-, ńʼdii-)?; (e) what does 33/13's "Pos. V, 3i obj. pron., something ?" mean, and is yináʼa- 765/9 alike?Print pp. 33, 44, 45, 49; Y&M grammar on prefix positions; Navajo unaccusativity literature. OPEN
T92a–d(new 2026-11-04, ~5am PT) From derivation: (a) is subject d a Voice-like spine head or a dP in Spec,Voice?; (b) is the postpositional head Appl or an incorporated P?; (c) is the word built by head movement or by post-syntactic merger?; (d) does Navajo need phrasal movement of argument dPs?Kari 1976; Rice 2000; Navajo binding/scope diagnostics. OPEN (a) and (c) answered in part on subject-d-site: Hale 2001 (22b) rules out a Voice-like spine head and a plain dP in a specifier for agentive deictic subjects; the spine needs no head movement, but the deictic d needs one displacement out of Spec,vP. HYPOTHESIS
T93a–d(new 2026-11-04, afternoon PT) From subject-d-site: (a) can the Voice-conditioned perfective allomorphy (Hale n. 7) be chosen in situ, or does it need Lowering?; (b) is the deictic d's move to the top of the spine post-syntactic or syntactic, and does it feed binding?; (c) does a deictic object d stay nonnuclear?; (d) is there any reason to merge a Navajo agent above Mode/Subj?DM locality sources (Embick & Noyer 2001); Hale 2000; forms with a deictic object and a deictic subject. OPEN Update 2026-11-04 evening: (a) answered in part on voice-allomorphy: in situ is enough under inside-out insertion with linear adjacency or spans (HYPOTHESIS); whether a single complex head is required is T94a. Update 2026-11-05 morning: (c) answered in part on deictic-order: with a 3i object and a 4th-person subject the object comes first (Y&M p. 123, 13/14 cells); the one reversed cell is a sound rule (HYPOTHESIS).
T94a–e(new 2026-11-04, evening PT) From voice-allomorphy: (a) does contextual allomorphy need one complex head (Embick 2010), which would bring Lowering back?; (b) why is the 2sg perfective the same with both classifier sets?; (c) is the 3rd-person z/s pair the same allomorphy or phonology?; (d) is the 1st dual-plural d-effect the same d (1du columns coincide)?; (e) does the perfective type depend on V across Voice (a span)?Embick 2010; Y&M print tables and grammar text on the d-effect; Kari 1976; stem entries. OPEN
T95a–e(new 2026-11-05, morning PT) From deictic-order: (a) is the ʼ zh → zh ʼ swap a stated rule of Navajo phonology?; (b) does a 4th-person object precede a deictic subject?; (c) what does 3a bizhʼdi- mean when the 3i is the subject?; (d) why does ʼa lose its vowel after da-/ná-?; (e) does any syntax see the swapped order?Y&M grammar on prefix phonology; Kari 1976; McDonough 2003; verb entries or Willie 1991 examples with both deictic prefixes. OPEN Update 2026-11-05 afternoon PT: (b) answered in part on space-d: the ho- object exponent (3s and 3a object, Y&M p. 458) precedes the 4th-person subject in 150 of 150 3a cells (hoji-, hozhdoo-) (HYPOTHESIS; data SOURCED); 3i subject and human 4th-person object still OPEN (T96a).
T97a–e(new 2026-11-05, evening PT) From qualifier-d-order: (a) why does p. 454 distributive future dahozhdiyoo- keep zh before diyoo while singular flips?; (b) do Y&M/Kari treat thematic di and future di as two heads or one fused piece?; (c) any other long hodii--type future 3a that does not flip?; (d) does between-di placement appear with 3i subject ʼ when thematic di is listed?; (e) is hozdoozis's z sibilant agreement (continues T96d)?Y&M print pp. 72, 451, 454, 541, 548 and grammar on the future / Pos. VI; Kari 1976; McDonough 2003. OPEN
T96a–e(new 2026-11-05, afternoon PT) From space-d: (a) a sourced verb with a human 4th-person object ho- and a 3i or 4th-person subject?; (b) do Y&M state the noun ban of hashbįʼ (p. 417) for other verbs?; (c) is ʼatiin in hałchxǫǫh (p. 411) the complement of ho- or a topic?; (d) why does 4th-person zh follow di in some hodii- future cells, and is hozdoozis's z sibilant agreement?; (e) why is hashtłʼó listed under ha-((1)) with no ho in its 3a forms?Y&M entries and print pp. 411–423, 72, 454, 541; Willie 1991; McDonough 2003. OPEN Update 2026-11-05 evening PT: (d) answered in part on qualifier-d-order: 6 of 46 ho+di+4th cells put zh after thematic di, all futures of long hodii- tables; short hodi- futures keep zh before di (SOURCED counts; HYPOTHESIS QD-5 refinement of step 5e). Sibilant half of (d) and (a)–(c), (e) still OPEN.
T89a–e(new 2026-11-03, ~12:46pm PT) From named-unnamed: (a) is the post-verbal bee-phrase an afterthought adjunct, and can it precede a 3i verb?; (b) can a Ø twin's named vehicle be a bee-phrase too?; (c) is 153/11's "(on an unspecified quadruped)" a translation slip?; (d) does ʼahééʼníłbą́ą́z (49/15), with chidí naabąąs as object, contain the 3i (possible counterexample for driver verbs)?; (e) do the 'a-((2)) vehicle verbs have Ø twins, and is yílwod one?Y&M sentence searches with bee/yee; paradigms of 49/15 and p. 123; the print edition. OPEN Update 2026-11-03 ~8:55pm PT: on driver-3i: (d) answered (HYPOTHESIS; data SOURCED): yes, 49/15 contains the 3i (prefix ʼahééʼa-, "Pos. IV, 3i obj. pron.", 49/14), but chidí is the subject of naabąąs in the object clause of bíhooshʼaah 'learn it', so this is not a counterexample; in the circle family 0 of 11 3i driver sentences have a vehicle noun as argument; (b) in part: łį́į́ʼ bee ʼahéénísdzá (53/1) has a named vehicle in a bee-phrase before a Ø verb. (a), (c), (e) still open.
T88a–e(new 2026-11-03, ~4:46am PT) From two-tables: (a) is p. 478's 3i future ʼadoo- right, given ʼeʼelyeed's future ʼiʼdoolwoł (311/10)?; (b) (reopens T85e for intransitives) are the 3i-headed ʼeʼe- intransitives 3i cells of 'a-((1)) bases or listed bases?; (c) does an unnamed-vehicle 3i verb ever take an overt vehicle noun?; (d) does łid ever occur with 3i ʼiʼííjool?; (e) what is the IV ʼa- of ʼeʼesooł 'swarm away'?Print p. 478; Y&M "3i person" headwords; wider Y&M sentence searches; stem entry *ł-zooł. OPEN Update 2026-11-03 ~12:46pm PT: on named-unnamed: (c) answered (HYPOTHESIS; data SOURCED): no, 0 of 32 sentences have the vehicle noun as an argument of a 3i verb; Y&M define twins as "named" vs "unspecified"; a coreferent vehicle appears only in a bee/yee phrase after the verb (772/3, 642/4); (d) in part: łid is overt with the doubled ʼiʼíí- perfective of ʼiʼiigeeh (468/4), still not with ʼiʼííjool. (a), (b), (e) still open.
T87a–e(new 2026-11-02, ~8:46pm PT) From listed-d (T85e): (a) is the single argument of the ʼeʼe- gas/sun/vehicle bases internal (unaccusative)? D's local listing check depends on it; (b) which verbs use p. 364's 3i perfective row ʼoʼoo-?; (c) does naʼajooł's perfective match a 3i row or row 3, and do other prefix classes behave like 'e'e-?; (d) under D, how does the stranded noun reach clause-initial position?; (e) with no overt noun, what values d's sort?Y&M 1987 entries and paradigm pages; Speas 1990; Willie 1991; Hale & Platero. OPEN Update 2026-11-03 ~4:46am PT: on two-tables: (b) answered (HYPOTHESIS; forms SOURCED): Ø/Ł 3i ʼoʼoo- appears only on transitive ʼeʼestsééh with an unexpressed agent (3i = PASS. A cell); (c) in part: 'a-((1)) (p. 478) has its own 3i row, identical to the ʼeʼe- forms, and LD-3/LD-9 are downgraded for intransitive bases (reopened as T88b); na- classes not reached; (e) in part: an unnamed vehicle takes the 3i with no noun (15/15), a named one the Ø twin (3/3). (a), (d) still open.
T86a–g(new 2026-11-02, ~2:30pm PT) From the revised essay Light determiners, analysis D (d takes a nominal complement; only d moves): (a) do bare nouns get definite and indefinite readings, and co-occur with demonstratives/quantifiers ([dP d DP] vs [dP d NP])?; (b) connectivity/locality of the noun (reconstruction, binding, islands)?; (c) exactly one noun per d, or stacked appositives?; (d) two quantified nouns, one per argument: true scope or restriction only? And is a noun with ʼa- read definite or indefinite?; (e) same behaviour on V, P and N hosts (the P object never goes silent, p-object-b)?; (f) can the noun be wh-moved, relativized or focused without the prefix?; (g) how does a subject d reach the word (Agree + defective D, or post-syntactic attachment)?Speas 1990; Baker 1996; Jelinek 1995; Willie 1991; Navajo texts and speaker judgments. OPEN
T85a–f(new 2026-11-02, ~2:30pm PT) From the essay Light determiners (Grant's conjecture; analyses A/B/C): (a) must a noun's features ever match the prefix (e.g. a non-indefinite noun blocked with a 3i base)?; (b) is any noun ever obligatory?; (c) standard PAH diagnostics (free noun order, VP-internal asymmetries, binding and quantifier scope of nouns), checked in Speas 1990 and Baker 1996 rather than assumed; (d) does any single base allow both 3i and Ø with the same noun (continues T83c, T84c)?; (e) is the 3i/Ø choice part of Y&M's verb theme entries?; (f) is the light d smaller than DP (pro-φP, Déchaine & Wiltschko 2002)?Speas 1990; Baker 1996; Y&M theme entries and paradigm pages; Déchaine & Wiltschko 2002 diagnostics. OPEN Update 2026-11-02 ~8:46pm PT: (e) answered on listed-d: yes, the 3i is in the listed prefix string or headword, not a person-cell choice, for the bases opened (HYPOTHESIS; data SOURCED); (d) partly: no same-noun alternation in the bases opened, no "optional 3i" in Y&M (silence only). (a), (b), (c), (f) still open. Update 2026-11-03 ~4:46am PT: the (e) answer is downgraded for intransitive bases on two-tables (the same forms are p. 478's 3i row; Y&M call ʼeʼelyeed "3i pers. forms"); still HYPOTHESIS for transitive p. 364 bases; reopened as T88b. Update 2026-11-05 afternoon PT: on space-d: (b) converse answered (SOURCED): Y&M p. 417 say hashbįʼ 'build a hogan' cannot take a noun object (*hooghan hashbįʼ); (a) in part (HYPOTHESIS): the same ho- allows naanish (p. 418), so the restriction is on the root, not a noun-prefix feature match. (c), (f) still open.
T84a–e(new 2026-11-02, ~1:00pm PT) From hang-subject: (a) is the 3i/Ø split in the 'hang' set partly aspect (si- neuter vs dah + progressive)?; (b) is dah part of the base that lists the 3i on *Øgoh?; (c) does any *Økał/*Økaad/*Øgish base take 3i with a gas noun?; (d) is neestin free of 3i and AR in its other modes?; (e) T83e continued: any productive rule for the 3i bases (motion vs handling stems, the sun as exception)?Y&M paradigm pages for the hanging bases; full reads of the saved kał/goh sets and a kaad/gish search; the nee- and ʼiʼii- prefix entries. OPEN
T83a–e(new 2026-11-02, ~5:00am PT) From rain-fire-subject: (a) is the areal ho- of the rain verbs the storm or the area rained on?; (b) is the 3i of the light-rain and blister bases 3i-CL or 3i-PRO, and does tó in tó ʼiiłtą́ behave like the gas noun?; (c) does any *łtįįh base take the 3i, or any *Øjooł base take ho- (would undo RW-13)?; (d) which ʼadi- is in ʼadikʼą́ą́h 'fire spreads'?; (e) is any productive rule left, e.g. 3i + di- 'fire' = light? Also: T82c answered as HYPOTHESIS (RW-10–RW-13); partial answers to T39d (Y&M p. 72 def 9: the ʼa- of ʼahoʼnii- is "Pos. Ib, away out of sight") and T39e (Y&M give rain verbs an areal 3s subject).Navajo texts; Willie on pleonastic ho-; Y&M stem index and grammar on ʼadi- and tó- prefixes. OPEN Update 2026-11-02 (~1:00pm PT): T83c answered on hang-subject (HN-9, HYPOTHESIS; data SOURCED): no base has both 3i and ho-, so RW-13 stands, and the listing unit is the base, not the stem. T83e partly answered (HN-11: the "formless/flowing stems take 3i" rule fails), carried on as T84e. T83a, b, d still open.
T82a–e(new 2026-11-01, ~1:00pm PT) From sun-subject: (a) is the 3i in the ł light verbs (chʼíʼdeełdlaad, ʼiʼdííłdláád) really the light in object position, with the sun or the person as causer (NF-14)?; (b) why does wind (níyol) take no 3i if [nat] licenses 3i-CL?; (c) do rain, water and fire verbs use the 3i (unread prefixes tóʼii-, tódaʼdii-, tanaʼa-)?; (d) are the sun-set verbs noun-less because the sun is recoverable or because they are lexicalized time expressions, and does any literal sun-motion 3i verb occur with jóhonaaʼéí?; (e) is the ʼa- of the thunder verbs the 3i (Y&M's own "?") or thematic? Also: T81b answered as HYPOTHESIS (NF-10): Y&M use the 3i for the sun on the handling stem *Øʼaah ("something (i.e. the sun)", p. 581 def 14, p. 704 def 5; SOURCED), for light, for melting snow and for a lightning strike, all without a noun (0/36 sentences). [cl: gas] becomes a sort value [nat] (NF-11; first half of T81d answered as HYPOTHESIS).Y&M paradigm and prefix entries; Navajo texts; Y&M grammar on Pos. IV vs V. OPEN Update 2026-11-02 (~5:00am PT): (c) answered as HYPOTHESIS on rain-fire-subject (RW-10): the rainstorm on *łtįįh takes the areal ho- (7 entries, 0/16 sentences with a rain noun), falling rain and hail, flowing water and fire take a plain third person with the noun, and the 3i appears only in light-rain and water-blister bases (SOURCED, Y&M pp. 38, 72, 390, 461, 527, 583, 630, 632, 635, 674, 707). [nat] is not sufficient (RW-11); the marker is listed per verb base (RW-13). (b) bears: wind joins the Ø nature verbs.
T81a–e(new 2026-11-01, ~4:50am PT) From gas-subject: (a) can a gas-set *Øjooł verb occur with no noun, and is it then 3i-PRO or 3i-CL?; (b) is 3i-CL limited to *Øjooł or used by other natural-force stems?; (c) why do the smoke-in-the-eyes verbs ʼajooł/ʼiijooł lack the 3i (expected ʼeʼe-)?; (d) is [cl: gas] a noun class or a sort value, and does the ranking treat a 3i-CL subject as definite?; (e) is the gas noun clause-internal or a topic or appositive, and D1 (VPISH Agree) or D2 (lexical class plus B matching)? Also: T80a answered as HYPOTHESIS (GM-8): in 9 Y&M *Øjooł entries for clouds, smoke, fog, rain and snowstorms, all 19 sentences have a noun with the 3i, and Y&M say the 3i "represents the substance" (p. 362 def 8; SOURCED). Wool and hay take Ø: five minimal pairs (GM-6). PX-11 is restated as PX-11′ (π or a valued class feature, GM-10). TI-13 resolved as HYPOTHESIS. Update 2026-11-01 (~1:00pm PT): (b) answered as HYPOTHESIS on sun-subject (NF-10: the sun, light, snow melting and lightning strikes also take 3i-CL); (d) first half answered as HYPOTHESIS (NF-11: a sort value [nat], not a noun class); (a) partly: noun-less 3i-CL exists for the sun set, still unknown for the gas set.Y&M grammar on 3i; Y&M paradigm pages and definition searches; Navajo texts on the box. OPEN
T80a–e(new 2026-10-31, ~9:30pm PT) From indef-subject: (a) níłtsą́ ʼadeezhjool has a bare noun beside a 3i subject (Y&M p. 16 def 15): linked (PX-11 wrong), attached otherwise, or thematic ʼa- (TI-13)?; (b) does a 3i subject change the classifier (naʼanish "deletion of l-classifier"), making the impersonal a derived voice?; (c) (replaces T77d) a human 3i object (Pos. IV "someone"), and does the ranking ever read 3i's sort?; (d) is bí- in bíʼnílcheeh a postposition or a direct object (IS-9/TI-10)?; (e) how do 3i ʼa- ('when one is a young girl', ʼídlį́) and the fourth person (IMP-4) divide generic 'one', and do Algonquian VAI impersonals match 3i on agentive intransitives?Y&M 1987 grammar (p. 67, p. 120); Neudorf 2000; Willie ch. 4, 9 and pp. 118–126 on images; Y&M paradigm pages; Wolvengrey 2011. OPEN Update 2026-11-01 (~4:50am PT): (a) answered as HYPOTHESIS on gas-subject: the noun is linked. Across the *Øjooł family, gas movers take 3i with an overt noun in 19/19 Y&M sentences, and Y&M say the 3i "does represent the substance functioning as the subject" (p. 362 def 8) and "a definite substance" (p. 401 def 9; SOURCED). PX-11 becomes PX-11′ (GM-10). (b)–(e) still OPEN.
T79a–d(new 2026-10-31, ~12:50pm PT) From x-on-third: (a) does Oxford's probe model cover X→3, given that a transitive X→3 with C5 Infl predicts -iko (XT-9)? Is Ojibwe X→3 also built on -aa?; (b) (= T78c, Navajo half) does INDF ʼadi- stay outside the 3rd/4th-person system?; (c) (= T78d, Navajo half) a sourced INDF form under a 'try'-type control predicate: can the unexpressed agent be the one who tries?; (d) is conjunct X→3 -iht segmentable, and does it contain a cognate of the PA/Kickapoo IMP -en?Oxford 2017/2019; Nichols 1980; Valentine 2001; Willie ch. 6, 9 on images; Young & Morgan 1987; Wolfart 1973; Dahlstrom 1991; Goddard 1979. OPEN
T78a–d(new 2026-10-31, ~4:46am PT) From impersonal-x: (a) Algonquian X ranks above 3 on the independent theme-sign hierarchy but lacks [PERS] (Oxford 2022 (23b)) and sits below animate in (22a): is an X→3 form direct or inverse? (bears on EI-8, T70e); (b) is Willie's "3s" for inanimate 'it' a real person value in Navajo (AX-10), and is there a sourced inanimate-on-1/2 form?; (c) can X be doubled by a nominal, and does Navajo INDF stay out of the 3rd/4th-person system as X stays out of obviation (Oxford 2018 fn. 13)?; (d) a Navajo test for a present but personless agent vs no agent (the Algonquian passive vs indefinite-actor dispute, Oxford 2022 fn. 15). Also: T76d answered (AX-8: X is in the paradigm as a cell but has no [PERS], SOURCED as Oxford's analysis; the parallel with ʼadi- is HYPOTHESIS AX-9).Goddard 1979; Ojibwe/Cree grammars; Goddard 1969, Dryer 1996; Willie ch. 2–3, 6, 9 on images. OPEN Update 2026-10-31 (~12:50pm PT): on x-on-third, (a) answered from Wolvengrey 2011 (Plains Cree): independent X→3 has the direct theme -ā, "formally identical" to 1/2→3 direct (p. 165, (112c)), the conjunct has -iht (neither direct nor inverse, Table 2.66), and no actor can ever be expressed ((106), (114)), so it is direct in form only (XT-8, SOURCED forms). AX-12 is dissolved: the X>3 rank restates this split (p. 163; XT-10, HYPOTHESIS). The Algonquian halves of (c) and (d) are answered: no doubling (XT-12), and kakwē- 'try' control shows the actor is present but unindexed (XT-6, XT-13). The Navajo halves continue as T79b and T79c. (b) is still OPEN.
T77a–d(new 2026-10-30, ~8:46pm PT) From absolutive-indef: (a) does any intransitive verb whose single argument is an agent take ʼa-, and is that participant animate (Willie p. 38 S2)?; (b) what does IMP refer to in (94) tsin baah haʼasʼnaʼ, where the tree is the postposition's object?; (c) is there a human indefinite intransitive subject (INDF in an ABS slot), against IS-10?; (d) is IMP's [−animate] read by the ranking (70)? (sharpens T76c). Also: T76a answered as HYPOTHESIS (IS-7, IS-8: ʼa- is the inanimate indefinite absolutive, Willie pp. 36, 38–39, 57–58; four of five statements say "inanimate").Young & Morgan 1987 (around p. 120); Willie 1990; Willie ch. 2 and ch. 4 on images. OPEN Update 2026-10-31 (~9:30pm PT): on indef-subject, from Young & Morgan 1987 (dictionary, first use on the site): (a) answered yes. Y&M's Pos. V "3i subj. pron., someone" ʼa- is on 'go', 'work', 'dream', 'be' with human referents (p. 303 defs 3, 21; p. 549 def 24; p. 555 def 17; p. 465 def 21; SOURCED), so Willie's S2 has examples and IS-8 is too narrow (TI-7, TI-8, HYPOTHESIS). (c) answered: human indefinite intransitive subjects are ʼa-, not ʼadi-, which Y&M keep for the Passive B agent ("some person", p. 178 def 3); IS-10′ keeps the case split and drops the sort split (TI-9, HYPOTHESIS). (b) answered as HYPOTHESIS: IMP in (94) is the 3i subject, the climber; the verb haashʼnééh is an intransitive with a P object (p. 434 def 8; TI-11). (d) restated as T80c. Also T33e answered (TI-12).
T76a–d(new 2026-10-30, ~12:46pm PT) From indef-person: (a) is intransitive ʼa- animate (Willie p. 38), inanimate (p. 39) or unrestricted?; (b) PX-11 (a bare noun links only to a slot with a person value) vs Willie's definiteness wording (p. 51): any counterexample?; (c) does an IMP object enter the animacy ranking (a bi- form with ʼa-)?; (d) Algonquian: is the impersonal actor X in a person paradigm or outside it? Also: T75a answered as HYPOTHESIS (PX-10: Willie pp. 221–222 list ʼadi-, ʼa and ho- "in addition to the person paradigms", and p. 86 ranges over person slots); T75d answered (all three personless, SOURCED; sort value SOURCED for ʼadi- and ho-, OPEN for ʼa-). Update 2026-10-30 ~8:46pm PT: (a) answered as HYPOTHESIS on absolutive-indef (inanimate, ABS); (c) sharpened as T77d; (b), (c), (d) still open. Update 2026-10-31 ~4:46am PT: (d) answered on impersonal-x (X is in the paradigm as a cell but has no [PERS], Oxford 2022; follow-ups T78a–d); (b), (c) still open.Willie ch. 3 on images; Willie 1990; Young & Morgan 1987; Oxford "Algonquian". OPEN
T75a–d(new 2026-10-30, ~4:46am PT) From passive-rank: (a) does p. 86 "only third person arguments can be indefinite" give INDF a positive [3], or only say indefinites are not 1st/2nd?; (b) any gloss of an agentive passive with a 3rd patient as "3.PASS" (chapter 2 style)?; (c) is INDF's [human] read by the ranking or only an interpretive restriction?; (d) do the other indefinite prefixes ('a, ho-) also lack person but carry animacy? Also: T74d answered as HYPOTHESIS (the ranking reads animacy; "PASS 3.3" is a paradigm label, AP-6/AP-7); T73d's direct evidence still OPEN. Update 2026-10-30 ~12:46pm PT: (a) answered as HYPOTHESIS and (d) answered (person SOURCED; ʼa-'s sort OPEN as T76a) on indef-person; (b), (c) still open.Willie ch. 3–4 and §2.3.2 on images; Willie 1990; Young & Morgan 1987; T70a. OPEN
T74a–e(new 2026-10-29, ~8:46pm PT) From agent-ext: (a) a diagnostic for EXT vs INT of a relative head when the other argument is 1/2/4 (restrictiveness cannot tell); (b) does p. 181 "Inverse relative clauses are also non-restrictive" cover (19), or only relatives binding the inverse EXT (p. 227)?; (c) p. 227 "subjects of Inverse constructions": inside the relative or the main-clause host?; (d) p. 50–51 class ʼadi apart from the persons vs the gloss "PASS.3.3": which does the ranking use (T73d)?; (e) are XA-4 (nouns) and HP-8 (REL) one rule?Willie 1990; Willie pp. 36–38 and ch. 3–4 on images; Young & Morgan 1987; a sourced noun inside a 1st/2nd-person relative. OPEN
T73a–e(new 2026-10-29, ~12:54pm PT) From person-head: (a) is the 3rd agent of a 3-on-1 relative EXT, as Willie p. 123 says for main clauses?; (b) a 1st/2nd-person relative with a noun on the third person (schematic [N siztsʼǫs=REL]): is the noun on the head, as (56) with 4th person?; (c) is FIX (no restrictive relative on an argument whose reference is already fixed) one condition or three?; (d) the case against EI-8 now rests on EI-10 (INDF on the animacy hierarchy): any direct evidence?; (e) Algonquian: with a 1st/2nd argument, does a relative or participle take its head from the only third person? Update 2026-10-29 ~8:46pm PT: (a) answered as HYPOTHESIS on agent-ext (XA-8: Willie p. 181 builds a relative from an independent clause, so p. 123 applies inside it); (d) one leg of EI-10 weakened by pp. 50–51, still OPEN (T74d). Also T72a partly answered: Willie p. 163 prints non-restrictive (5) without commas, so a comma-less gloss is no evidence of a restrictive reading.OPEN
T72a–d(new 2026-10-29, ~4:53am PT) From external-head: (a) Willie (20b) binds the definite inverse external patient with no commas in the gloss, but *(24b) is excluded by definiteness: is (20b) non-restrictive, or does the external argument's own noun lift the clash?; (b) an agentive passive with one noun on the patient, [N biʼdistał=yę́ę] (schematic): non-restrictive only, as VC-3 predicts?; (c) Platero 1982's two-reading speakers: is (16) grammar or preference?; (d) does the noun-count dependence of the relative head hold with 1st/2nd-person arguments (Willie pp. 170–175)? Update 2026-10-29 ~12:54pm PT: (d) answered on person-head: no. With one third person the head is that third person, agent or patient ((41a), (41b), (45)–(57), SOURCED readings). The noun-count grid holds only for two third persons (HYPOTHESIS). (a)–(c) still open. (a) is now sharpened by FIX (HP-11).OPEN
T71a–e(new 2026-10-28, ~8:46pm PT) From third-voice: (a) can the indefinite agent ʼadi- head a relative (EI-8′ says yes by Willie (16), the ban on adjoined nouns and EI-9 say no)?; (b) is there any inverse relative whose head is the external patient (Willie (16) in the S/INT sense predicts none)?; (c) which classifier biʼdistał carries, and is the intransitivity in the classifier (Willie p. 36) or in Hale's ergative d?; (d) any diagnostic of Absolutive vs Accusative apart from prefix shape?; (e) the inverse agent: "internal ... not the subject" (p. 72) or "the transitive subject of the relative" (p. 160)? Update 2026-10-29 ~4:53am PT: (b) answered on external-head: yes, Willie (20b) with (21), so (16) is the default for relatives without nouns (SOURCED data; HYPOTHESIS scope). (e) partly answered there: internal in ch. 6 too (HYPOTHESIS). (a), (c), (d) still open.OPEN
T70a–f(new 2026-10-28, ~12:46pm PT) From ergative-indefinite: (a) a sourced agentive passive with a non-human 3rd patient (EI-8 predicts it is bad like Willie *(72b.4), EI-9 predicts it is fine); (b) are ʼadi- forms with 1st/2nd objects Direct or Inverse, and what does "Ergative" mean there?; (c) the case of the agentive-passive patient (Absolutive p. 100 vs Accusative p. 72); (d) a form that separates FS-8a from FS-8b (narrative 4th with an overt noun + 3rd object); (e) Algonquian X>3: direct or inverse?; (f) = T69f restated: an impersonal (generic) ji- with a 3rd object.OPEN Update 2026-10-28 (~8:46pm PT): on third-voice, (b) answered: with a 1st/2nd patient the agentive passive is neither Inverse (not all 3rd, Willie p. 86) nor Direct (its agent is Ergative = internal, whereas a Direct 3rd agent with a 1st patient is "external Nominative", p. 123), so it is a third voice (VC-13, HYPOTHESIS). (c) answered: the patient is Absolutive in both of Willie's senses (it is the "subject" of the intransitive-classifier passive and heads a relative under her (16), pp. 152, 160), not Accusative (VC-14, HYPOTHESIS). Consequence: EI-8 survives only as an intransitive inverse (EI-8′). (a), (d), (e), (f) still OPEN. Update 2026-10-31 (~12:50pm PT): (e) answered on x-on-third: Plains Cree X→3 is direct in form but intransitive (no actor can be expressed, Wolvengrey 2011 (114)), and an overt indefinite 'someone' acting on a 3rd is inverse ((113)), so the analogy against EI-8 loses its basis (XT-11, HYPOTHESIS). (a), (d), (f) still OPEN.
T69a–g(new 2026-10-28, ~4:46am PT) From indefinite-agent-b: (a) a P object under an agentive-passive subject; (b) what handout (22)'s "indefinite" Ø trigger means, and whether any transitive ʼ-d- form shows Ø; (c) is the bi- of biʼdistał the elsewhere b- or an inverse-type bi- (scope, word order)?; (d) do Willie p. 115 náhodiiʼnil / nábiʼdiiʼnil contain ʼ-d-?; (e) is the indefinite agent on the animacy ranking (FS-8b)?; (f) impersonal ji- with a 3rd object: is Ø due to the morpheme or definiteness?; (g) Algonquian unspecified-actor (X) object agreement (merges T68f).OPEN Update 2026-10-28 (~12:46pm PT): on ergative-indefinite, (e) answered: by Willie's own criteria ("definite, human"; only 3rd persons are indefinite) the indefinite agent is on the ranking, so FS-8b survives and picks out the same class as FS-8a (HYPOTHESIS). (g) answered: Algonquian X is ranked 1/2>X>3 and X>1 shows the inverse theme sign with an overt object, like ʼ-d- (SOURCED for Meskwaki; the comparison is a HYPOTHESIS). (c) partly answered: Willie's "Ergative" (internal agent) vs Hale's (subject) maps onto IX-8 vs IX-6, and the deciding test is now T70a. (f) refined as T70f.
T68a–f(new 2026-10-27, ~8:46pm PT) From fourth-subject-p: (a) a second 4th-subject + 3rd P-object token outside the quotative; (b) Hale 1997 (zero AGRs, not 3rd in obviation) vs Hale 2001 (no person feature): which derives no yi-?; (c) is H checked against the P's own clause subject (needs 4th inner subject, 3rd matrix)?; (d) is the object under ji- a true Ø or an absorbed yi-?; (e) 4th-person P object ha- with a 3rd subject; (f) Algonquian unspecified-actor forms and obviation. Update 2026-10-28 (~4:46am PT): (b) answered on indefinite-agent-b. The agentive passive ('someone' as doer) gives bi- where ji- gives Ø, so a Ø keyed to "position V filled" fails, and Hale 1997 fn. 9 + elsewhere b- (FS-8a) derives the data (HYPOTHESIS). (d) partly answered: object prefixes do surface before the deictic slot (weak support for a true Ø, HYPOTHESIS).OPEN
T67a–e(new 2026-10-27, ~12:47pm PT) From p-object-b: (a) any sourced Ø P object under 'I'/'you'?; (b) 4th-person (ji-) subject with a 3rd P object; (c) are proclitic Ps like béé- inside the phonological word?; (d) a sourced 'you' clause (all 8 have 'I'); (e) an Algonquian counterpart of the P/E split in silence. Needs: Y&M 1987 paradigms; Kari 1976; McDonough 1990; Willie ch. 4–5. [Update 2026-10-27 ~8:46pm PT: (b) answered on fourth-subject-p. Under a 4th-person subject the P object is b- (Willie (11) bił vs (12) yił, 1 token) and the verb's 3rd-person object shows no yi- (11 tokens), as Hale 1997 fn. 9/17 predict. Also bears on T15b: bił in (11) is b- because H fails under the 4th person.]OPEN
T66a–f(new 2026-10-27, ~11:29am PT) From the verb guide: (a) Y&M stem sets by mode; (b) Routledge Table 3 TI/TA glosses; (c) does bi- reverse binding; (d) prefix–classificatory-stem agreement (T40h); (e) are phrasal verbs listed as lexemes; (f) a Navajo counterpart to LeSourd 2006.
T65a–e(new 2026-10-27, ~4:46am PT) From causee bi- under 'I': (a) is causee b- under 'I'/'you' obligatory (H&P "is overt") or optional (Saxon fn. 12 "can be used")? Is there any Ø-causee form?; (b) 4th-person causers: CB-5 predicts b-; (c) more inanimate causees under 'I' besides (43b) the car, against Saxon's "animate"; (d) Saxon (42) b-b- 'We showed him how to drive a truck': which b- is the causee, and is it compatible with HG-9?; (e) do non-causative incorporated Ps also keep b- (never Ø) under 'I'/'you'? [Update 2026-10-27 ~12:47pm PT: (e) answered on p-object-b: yes in all sourced data (9/9 b-); ordinary Ps are proclitics, not incorporated (Hale p. 53).] Needs: Y&M 1987 causative paradigms and pp. 223, 267; Willie 1991 ch. 3–4; a native-speaker judgment for (a).OPEN
T64a–f(new 2026-10-26, ~8:46pm PT) From Hale's "general" b-: (a) the printed Hale 2000 pp. 70–71 vs hale023 PDF pp. 16–18 (footnote 11 vs 12); (b) can a thing-subject raise and license yi-? (Saxon's rainbow pair, Y&M 1987:225, = T61c); (c) (c23) yikʼi vs (32) báhátis: subject height or PP attachment?; (d) a reordered Perkins (8c) with 'bike' before 'girl' separates the order-only and slot readings; (e) Willie (85)/(86) verb yoolóós (page image) vs yoołóós printed on pp-by-verb, postposition-units and procliticized-p; (f) is an incorporated P's agreement under 'I'/'you' always b-, never fn. 9's Ø? [Update 2026-10-27 ~4:46am PT: answered for causatives on causee-bi. All 10 sourced causee forms with a 1st-person subject have b- and none has Ø (SOURCED), and Hale & Platero 1996 state the pattern themselves. Ø is limited to E's own AGRo (CB-5, HYPOTHESIS). Whether it is "always" so, and other incorporated Ps: T65a, T65e.] Needs: Hale 2000 printed; Y&M 1987 pp. 225 and causative paradigms; Perkins 1978 pp. 55, 116, 136; Creamer 1974; Willie pp. 95–96 images.OPEN
T63a–d(new 2026-10-26, ~12:46pm PT) From three third persons on Aissen's 1997 tier: (a) the both-obviative test: with a topical third person elsewhere, does the prefix in a clause of two non-topical 3rd persons follow which one is more prominent (R5) or always come out YI (Aissen 2000)?; (b) does any Navajo form show Obv1 vs Obv2 directly (a double possessive, like the Blackfoot further obviative)?; (c) per-word comparison (RO-8): is there any clause where the verb's and a postposition's prefix must share one value? [Update 2026-10-26 ~8:46pm PT: partly answered on hale-general-b §7. No forced shared value was found. Hale's causative (49)–(51) shows a limit on number instead: one referential object agreement per verb word, and an incorporated P's agreement takes it (SOURCED). The word is the domain (HYPOTHESIS), consistent with RO-8]; (d) Thompson 1989/1994 first-hand: a relative rule? Needs: Navajo text passages; a sourced double possessive; Perkins 1978; Willie 1991 postposition chapters; Thompson.OPEN
T62a–e(new 2026-10-26, ~4:46am PT) From one "main" third person per clause?: (a) clean evidence for "one bi- per clause" with a human subject once animacy is controlled (Perkins (8c) is confounded under subject-relative yi-, OP-8); (b) the rider effect: is the thing-object bi- in Y&M (38)/(40) caused by the human rider? (a minimal pair with/without 'with him'); (c) Creamer (c23) yikʼi vs Perkins (32) báhátis: speaker variation, the postposition, or context? (Perkins 1978 pp. 55, 136 first-hand); (d) Aissen 1997 on spans with three 3rd persons [Update 2026-10-26 ~12:46pm PT: answered on ranked-obviatives. Aissen 1997 allows one proximate and several ranked obviatives (§2, §3.5 (34)), and her Algonquian direct/inverse compares subject and primary object ((12)–(15), fn. 10). She has no Navajo YI/BI rule (p. 736). R4 = that comparison applied to Navajo (R5, HYPOTHESIS); see T63a–d]; (e) Hale 2000:70–71 first-hand: is the "general third person object prefix b-" the subject-relative idea? [Update 2026-10-26 ~8:46pm PT: answered on hale-general-b from Hale's 1997 workshop text (hale023), which contains Saxon's "Hale 2000:70–71" quotations verbatim (identity HYPOTHESIS). Yes in shape: y- only when the object is in the c-command domain of a clausemate "true" 3rd-person subject, b- general (SOURCED). The relation is c-command, not rank; equal to R5 under BR (HYPOTHESIS). See T64a–f] Needs: Y&M 1987 pp. 151, 227, 475; Perkins 1978; Aissen 1997; Hale 2000; a native-speaker judgment for (a).OPEN
T61a–f(new 2026-10-25, ~8:46pm PT) From two ways to put the object first: (a) the binding datum T60c still lacks: does a bi- (Spec,DiscP) patient bind a reciprocal or bound pronoun in the doer's phrase, or show weak crossover, while a Focus-Moved (yi-) object does not?; (b) does Focus Movement reconstruct for binding as well as for the prefix, and is non-wh Focus Movement iterable and unbounded?; (c) Saxon fn. 9 rainbow pair (Y&M 1987:225): yi- with a thing-subject, and agency (Willie 2000:365) vs number as the licensor; (d) number feature (Saxon) vs Proto-Agent strength: which keeps a subject in VP?; (e) does Aissen 2000 address two bi-s with inanimate subjects? [Update 2026-10-26 ~4:46am PT: answered on one-proximate. Aissen does not discuss such clauses, but her fn. 15 ("exactly one proximate") with fn. 5 (BI/YI are nonsubject pronouns) predicts one yi- in Y&M (38)/(39), so they are contradicting data (SX-8 confirmed against her text). A subject-relative yi- (R4) fits; see T62a–e.]; (f) Rice & Saxon 1994, Hale 2000:70–71, the WSCLA 6 version and Saxon 1984a/1986/1995 first-hand. [Update 2026-10-26 ~8:46pm PT: Hale 2000:70–71 partly: its text is read in the 1997 workshop version (hale023) on hale-general-b; the printed pages are still unopened (T64a). The others remain unopened.]OPEN
T60a–f(new 2026-10-25, ~12:46pm PT) From what comes first decides: (a) a bracket-forcing singular test: under bi-, is coreference out when the owner's name is forced inside a singular doer's phrase?; (b) Platero (3.18) ✓ vs Young & Morgan (29) * on a coreferent owner inside the doer's phrase under yi-: repeated name vs null owner, or speakers?; (c) is the first-position null patient under bi- in an A-position (deep) or a topic (A′)? (reciprocal / weak crossover) [Update 2026-10-25 ~8:46pm PT: partly answered on two-osv. Saxon 2003 treats only Focus Movement (yi- kept) as A-bar, and her bi- slot Spec,DiscP behaves A-like (HYPOTHESIS), which favours option (ii). There is still no binding datum, so this continues as T61a]; (d) owner bi- disjoint from a 3rd-person doer (Platero (4.23)) vs Horseherder's owner yi-: speaker variation?; (e) do PP command boundaries hold under bi-?; (f) Platero p. 133 vs HJP 1976:403 on relatives.OPEN
T59a–f(new 2026-10-25, ~4:46am PT) From one "his", two third persons: (a) W&J (70) with a singular possessum ('John's son saw him', bi-): does coreference survive? (separates plural doer from 'all' / name-inside); (b) the D2 sentence [N_P] [yi-N_A] bi-V: good (deep bi-) or bad (shallow bi- / ranking)?; (c) does a dual doer block owner yi- like a plural one (Horseherder (6b) p. 47 has no da-)?; (d) a quantified A′ antecedent for owner yi- (= T57c); (e) Saxon 1984a/1986/1995 and Aoun 1985 first-hand [Update 2026-10-25 ~8:46pm PT: partly. Saxon 2003, a later Navajo paper of hers, was opened (two-osv). Its rule, yi- only under c-command by a 3rd-person subject in Spec,NumP, fits Horseherder's report, but its fn. 1 sets possessives aside. 1984a/1986/1995 are still unopened]; (f) does a fourth-person doer count as a disjoint third person for owner yi-? Update 2026-10-25 ~12:46pm PT (possessor-command): (a) partly answered: Platero 1978 (1.63a) Ashkii bizhéʼé biiłtsą́ 'The boy was seen by his father' is good with coreference, but on the name-outside parse (SOURCED). His precede-and-command rule plus his IGR predict (70b)'s star with no plural condition, and 'all' only forces the name inside (HYPOTHESIS); name-inside singular test moved to T60a.
T58a–f(new 2026-10-24, ~8:46pm PT) From where the owner's name sits: (a) why W&J (70b) 'all John'si children saw himi' is bad under bi- when (30) is good: name inside the doer's phrase, 'all', or source?; (b) Platero 1982 / Hale, Jeanne & Platero 1977 first-hand; (c) inanimate-owner test (Tzotzil (75) analogue); (d) Thompson 1989/1994; (e) is Cree (19)'s name inside the possessed phrase?; (f) operator binding (merged with T57c).OPEN Update 2026-10-25 ~4:46am PT (possessor-yi): (f) partly answered: Horseherder 1998 binds owner yi- from an A′ position (possessor or discourse topic), but tests no quantifiers (HYPOTHESIS); (a) partly answered: a new lean, the plural doer, since her (6b) p. 47 'they roped her horse for her' is * with a plural doer and no name inside the doer's phrase, where ranking predicts ✓ (HYPOTHESIS).
T57a–d(new 2026-10-24, ~12:46pm PT) From who can "his" point back to under bi-?: (a) with a fourth-person antecedent or possessor (outside obviation), does the possessor restriction vanish or stay?; (b) Young & Morgan 1980:171–172 first-hand; (c) a quantified antecedent ('every boy … his horse') under yi- vs bi-; (d) the Algonquian possessor constraint first-hand (Rhodes 1993; Aissen 1997, needs OCR). Update 2026-10-24 ~8:46pm PT (possessor-gap): (d) answered: Aissen 1997 read (OCR); the Algonquian gap uses the 3′→3 inverse, but she derives it by ranking with no movement, and Tzotzil uses a passive, so it is not a deep-inverse test (HYPOTHESIS); Rhodes 1993 still not opened. (a) partly answered: with a 4th-person owner no yi-/bi- choice arises, and coreference returns (W&J (77)), confounded with the name's position.OPEN
T56a–c(new 2026-10-24, ~4:46am PT) From why did Hale move da- up?: (a) a Navajo causative with da-: must the causer be plural, or can the causees alone be? (separates Pl(L2) from Pl(H)/Pl(L1)); (b) under VPISH, do Hale's narrow-scope low adjuncts need subject raising to Spec,M/SP or a site below Spec,v?; (c) does Hale or Rice address the clash between Pv-outermost trees and leftward scope (PL-R2)?OPEN
T55a–d(new 2026-10-23, ~8:46pm PT) From how high is da-?: (a) Pl(H) at QualP or Pl(L) at vP: a diagnostic that separates them under VPISH; (b) does M/S agree with EXT or ρA in the inverse (= T10b, T30b; decides whether Pl(H) keeps DIST-6)?; (c) Rice 2000:80–81 and Rice's scope principles first-hand; (d) Hale's reason for moving Pl from VP (2001) to QualP (2003). Update 2026-10-24 ~4:46am PT (pl-move-reason): (d) answered: no Pl-specific reason, Pl moves with the adjunct class (SOURCED); (a) narrowed: prefix order ruled out, causative test proposed (T56a).OPEN (a–c); SOURCED (d)
T54a–d(new 2026-10-23, ~12:46pm PT) From can da- count the patient?: (a) can a bare EXT of three or more stand under bi- without da-? (b) is Pl merged above the inverse landing site (= T3, T30c)? (c) does the second probe ever apply in the direct with a fronted patient? (d) is shizhéʼé (64b) definite by itself? Update 2026-10-23 ~8:46pm PT: (b) answered in part on how high is da-?: Hale 2003 puts Navajo Pl at QualP or vP (SOURCED), which is above an outer-Spec,v landing site (HYPOTHESIS); Hale 2001's VP-internal Pl would not support PL-2H.OPEN (a, c, d); HYPOTHESIS (b)
T53a–f(new 2026-10-23, ~4:46am PT) From deep or shallow inverse?: (a) binding/scope under bi-; (b) is RCC hard, soft or speaker-variable?; (c) does RCC need a theme exemption (INV-5)?; (d) can da- count the raised patient alone?; (e) =go chains, Willie (9), damaged subscripts; (f) Platero 1982's two-reading speakers. Update 2026-10-23 ~12:46pm PT: (d) closed, negative, on can da- count the patient? (Willie (64a), (65b), (67b), (68)). (b) Sandoval & Jelinek pp. 345–352 re-opened: nothing on rank strength. Update 2026-10-24 ~12:46pm PT (possessor-binding): (a) partly answered: possessor coreference data (Y&M 1980 via Aissen 2000) fit a deep bi- (HYPOTHESIS), but obviation and word order predict them too, and there is no quantifier, so still open (T57a–d).OPEN (a–c, e–f); SOURCED (d: no)
T52a–d(new 2026-10-23, early morning) From the Algonquian inverse: (a) does any Algonquian inverse respond to a patient proto-property? (b) Blackfoot binding in sentience-ranked direct forms; (c) Klaiman/Givón typology of inverse voice (not opened); (d) is Navajo bi- a deep inverse by Oxford 2023's tests (= T10a/T11c)? Update 2026-10-23 ~4:46am PT: (d) answered as far as the data go on deep or shallow inverse?: deep by contrast, linking, relativization and definiteness (HYPOTHESIS); binding still open (T53a).OPEN (a–c); HYPOTHESIS (d)
T51a–g(new 2026-10-22, 8:46pm) After T50a: (a) NLA 92's subject (gloss “3-1” vs translation ‘my grandma’); (b) a sourced yee ‘with it’ (gap vs unranked; in a direct 3/3 sentence it refutes P1); (c) -tsʼą́ą́ʼ in a 3/3 sentence; (d) X3 with two third persons; (e) -ł with a thing in a 3/3 sentence (IP-7 a vs b), and NLA entry 323 (-kʼee, not opened); (f) more two-P sequences for IP-8; (g) bééhózin vs béé hózin, and Creamer p. 34's elephant name. Status of earlier items: T50a done for -ł, -ee (open for -tsʼą́ą́ʼ); T49b, T49c done; T49d first test passed (weak); T50b no counterexample; T50f → IP-8. T48a, T46h, T49a not attempted (NLA down).Instrument and voice §8
T50a–h(new 2026-10-22, 12:46pm) After T49f/T49h: (a) a yi-/bi- pair, or its absence, for -ee Instrumental, -tsʼą́ą́ʼ, -ł (decides IP-7 animacy vs relation); (b) a noun between a ranked PP (yá, yaa, yichʼįʼ) and W (predicted impossible); (c) a focus noun between an unranked source PP and W (predicted possible); (d) examples for -ghá, -láah; (e) is yi-/bi- on a P the voice or obviation (Willie p. 96)? (f) the order of two PPs inside QualP; (g) rewrite the clause spine's MVS-as-D1 text to CYC v2; (h) the beneficiary's site: VP-adjoined vs an applicative headPostpositions by verb §9; NLA lexicons (live site down); Young & Morgan
T49a–h(new 2026-10-22, 4:46am) After T49 (NLA (9)–(11b) bee dibé síníloh vs Willie's uninterruptible MVS): (a) a noun between a PP whose object is a verb participant and the verb (decides IP-1 vs a low-noun or low-focus analysis; would count against T-MID); (b) the object noun before an instrument PP; (c) an instrument sentence with two 3rd persons (yi-/bi-); (d) does the instrument's P alternate yee/bee with the voice; (e) PV-2: -ísh on dah/yah after a PP (predicted impossible), ka-type preverbs before a PP; (f) Willie's (28a–i) verb by verb; (g) Speas 1990 and Barss, Hale, Perkins & Speas 1989 on PP vs object noun; (h) rewrite the CYC rule text with D1 = W and C1′Noun between P and verb §11
T47, T48a–h(new 2026-10-21, 8:46pm) T47: is the distance phrase the incremental theme in motion events (Grant)? Theory: supported (Dowty, Krifka, HKL). After T47: (a) a sourced sentence with a vehicle and a distance (‘I came twenty miles by car’); (b) the ‘but didn't get there / didn't finish’ judgments T1–T5 (no source prints them); (c) can ni- “cess” and a distance phrase co-occur, and must they match? (d) Midgette 1996 and Smith 1996 on telic morphology; (e) page images of Smith, Perkins & Fernald 2007 (28b), (30a); (f) the segmentation of N1's distance phrase and of tsin sitą́ ‘mile’; (g) the subject of N6 shigáál deiz tʼiʼ, and N10 chʼééh dé yá a go; (h) Tenny 1994 and Levin & Rappaport Hovav in the originalDistance as measure §11
T46a–h(new 2026-10-21, 12:46pm) After T45f: (a) S12 shił ʼadees wod “1-with 1-drive.F”: rider = subject, or a gloss slip? (b) S13 bił yah ʼiʼííbą́ą́z: a pusher or the wheeled thing as subject? (c) productivity proper: a new vehicle or motion stem in the pattern (d) a movement stem with the traveller as subject and -ił holding the vehicle (would break TP-4) (e) which linking rule picks ‘lightning’ as subject in S2 (f) are dah/yah preverbs in Hale's sense; can -ísh follow them; any clause-level word between a PP and dah? (g) how Young & Morgan list ʼoolwoł ‘drive’ (blocked) (h) the four undisplayed “transportation” entries (re-open the live site)The -ił travel pattern §10
T45a–h(new 2026-10-21, 4:46am) After T43e: (a) -ísh on a postposition normally written onto the verb (e.g. the béé of bééhózin): possible, and does it then stand apart? (b) one source/speaker writing the same grammatical P both ways: free or conditioned? (c) do Young & Morgan list the idiom pairs (-ídin + ‘be’, -yaa + hazlį́į́ʼ, -chʼįʼ + sidziil, -ká + ‘help’) as units? (blocked) (d) any word printed between a grammatical PP and its verb (falsifies PC-2)? (12:46pm: only the preverbs dah, yah, which belong to the verb word, TP-6; rest → T46f) (e) can -ísh split a causative like Hale (41b) bi-di-y-é-saʼ? (predicted no) (f) is the -ił “transportation” pattern productive? (answered 2026-10-21, 12:46pm: a pattern built from the parts; strict productivity → T46c) (g) a printed attached grammatical P in Willie 1991 itself; (h) where doo … da, daatsʼí, ʼayóo sit in the spineAttached or separate postpositions §10
T44a–g(new 2026-10-20, 8:46pm) After T43f: (a) a shifted =sh with a first phrase of two or more words: after the first word or the whole phrase? (b) the unshifted inverse version of Willie *(17): fine, as QC-1 + QC-5 predict? (c) a yes/no question with focus fronting under yi-: is QC-4 inverse-only? (d) why the inverse freeze holds only in questions ((74a) is a fine statement); (e) is the =sh shift Move (a copy) or a spell-out choice? (f) the NLA's “other places” for -sh/lá, and topic -shąʼ vs =sh; (g) =sh and -ísh in one clause?The two question endings compared §9
T43a–h(new 2026-10-20, 12:46pm) After T41f: (a) a form with two or more disjunct prefixes plus -ísh: where can it go? (b) any -ísh/=sh inside a verb word (falsifies EN-1′)? (c) can yisdá or another preverb Hale prints inside the word host -ísh; is an unsplit ka… form possible (B1 vs B2)? (d) why NLA (5) neésh and (7) hólǫ́nísh are “?” while (11b) is fine; (e) can a procliticized postposition (W&J p. 270) host -ísh, and is D1 the verb word or the whole MVS? (f) polar -ísh vs content =sh: first word or first constituent; does -ísh show the inverse restriction? (g) the NLA's focus underlining in (10)–(11); (h) the ¡/¶ key applied to Hale's other broken-font forms. Status 2026-10-20, 8:46pm: (f) answered except for the shifted case, now T44a (the two question endings compared); the rest unchanged. Status 2026-10-21, 4:46am: (e) half answered: D1 = the verb word W, the MVS is an unsealed phrase (PC-1); the host half is now T45a/T45e (attached or separate postpositions)Where the yes/no enclitic -ísh can go §10
T42a–h(new 2026-10-20, 4:46am) After checking Chomsky 2000/2001: (a) PIC1 vs PIC2: does M/S ever spell the object's person when the Spec,v* slot is not a matching goal? (b) is the fourth person a person (φ) value or a separate deictic feature (decides whether MA-3a is possible)? (c) any sign that Navajo slots carry an activating (case-like) feature, or that a P-agreed slot is still visible to M/S (EP-5 vs MA-5a; joins T41c)? (d) no ‘start to’ form with a placeholder ’ plus a separately spelled participant or definite object (EP-3 prediction); (e) a ‘start to’ verb with two slots below v: which is bumped (prediction: the closer)? (f) is PINCH's agreement φ-complete (number on the bumped prefix)? is the placeholder there-type or it-type? (g) is INCH's requirement obligatory in every mode and aspect? (h) Chomsky 2004, from the EPP-feature to the edge featureThe older definitions behind “edge feature” §9
T41a–h(new 2026-10-19, 8:46pm; T41e answered, T41h partly, T41c sharpened, 2026-10-20 4:46am; T41f answered 2026-10-20 12:46pm: first-constituent enclitic, but the n. 12 split is between words → word-internal test OPEN, T43a/b) After checking the mid-level terms against Rudin, Hiraiwa and Chomsky: (a) SH-4 vs Chomsky's selection condition, which needs a form with DETR, a separate qualifier and a deictic subject; (b) the deictic at the top: Agree or edge-feature movement (any A′ property)? (c) the PP as a phase or as an inactive goal (an active goal below a PP); (d) Grewendorf 2001 full text; (e) Chomsky 2000/2001 EPP definitions; (f) Rudin's clitic test on a legible two-disjunct form with the polar enclitic; (g) does da- ever count a plural that M/S does not spell? (h) is ‘go’ in W&J (50b) built on a phasal v*?The borrowed terms checked against their originals §9
T40a–j(new 2026-10-19, 1:52pm; T40b answered, T40c conditionally, T40d narrowed, T40g partly, 8:46pm; T40a closed into T43a, 2026-10-20 12:46pm) After sorting word-internal facts into Merge and Agree: Rudin's clitic/indivisibility test on disjunct prefixes with the polar enclitic; open Rudin 1988, Hiraiwa 2001, Chomsky 2008, Grewendorf 2001; feature inheritance with DETR on top; Multiple Agree vs contextual allomorphy for the object prefix; U-1 vs U-2; the PP edge under the inverse; is vP a phase in the word; stem–prefix agreement (Grant's notes §4); a template position with two exponents; Person Case Constraint gapsMerge and Agree inside the word §10
T38a–h(new 2026-10-19, 4:46am) After T37b: an unaccusative ‘start to’ form with a disjunct prefix: is the participant's prefix first, like the oblique bi- of handout (23a) Bináádaʼniilghal, or next to ʼnii- (the handout's eccentric agreement)? Is the ’ of ʼnii- a subject (position V) or an object (IV) in the unaccusative case? Is the third person in Mode constant across persons? Does the ‘start to’ test sort the whole H&P/LINK-1 sample as the causative does (joins T4)? Is the inchoative ’ role-less (T17)? Why is the oblique merged above the disjuncts (joins T29f)? Hale, Munro & Platero 2000 in their own words; a 4th-person subject with ʼnii- (OBL ⌢ ’ ⌢ j ⌢ nii predicted). (2026-10-19, 12:46pm: T38a and T38b answered; T38d, T38e, T38f, T38g partly answered; T38c narrowed; T38h open; HMP page §11.)The ‘start to’ test §11; Young & Morgan 1987:187–188, 217 (not opened); Hale, Munro & Platero 2000 (not on the box)
T37a–h(new 2026-10-18, 8:46pm) After T35c/T36g: does a 4th-person gamma perfective show Hale's 3rd-person [w] (Willie joolyé, jooʼash)? Does a labile intransitive's slot raise to Spec,v inside the word, or is the portmanteau just Agree with the closest slot? Indefinite subject + 3rd-person object outside the passive: Ø or b-? Areal ho- with an object prefix; da- + ji- forms; Hupa/Ahtna deictic-before-object; clean copies of Hale 1997 (16), (20); restate the fourth-person pages with δ. (T37b partly answered 2026-10-19, 4:46am: the inchoative bleeds a low participant's subjecthood; U-2 favoured on economy, not decided; the ‘start to’ test.)Subject on the head §11; Young & Morgan 1987 (not opened); Golla 1970, Kari 1989 (not opened)
T36a–h(new 2026-10-18, 12:46pm) After T35b: Young & Morgan's own position for the reciprocal ʼahi- and the reflexive's di-? When does the reflexive lack ʼá-? Why does the reciprocal lack di-? Does ʼá- stay in place beside a preverb or oblique (object pronounced high vs binder)? Scope of ʼá- vs da- (Rice)? Saxon 1986's empty anaphoric pronominal vs role removal? Is the deictic subject on a head above di- (WF-6)? (T36g done 2026-10-18, 8:46pm: adjoined to DETR, subject on the head) Is Willie Ch. 3 (4) ʼádílcha a reflexive over a causative?Reflexive vs reciprocal §11; Young & Morgan 1987; Saxon 1986; Rice 2000; Kibrik 1996, 2001a
T35a–i(new 2026-10-18, 4:46am) After T27o: is the yi- of Willie (4a) yi-sh-kʼá, (10) yiiłtsą́, (12) yiichʼid the 3rd-person object, against her p. 28 and Hale 1997 n. 9? Why does the reflexive precede da- ((14) ʼádadiidleesh) but the reciprocal follow it ((18) daʼahiidleesh)? (T35b done 12:46pm: a two-piece reflexive, reflexive vs reciprocal.) Can “subject exponents on a head” (WF-6) be derived? Do Jelinek 1990 or Hale, Jelinek & Willie 1991 give a tree? Does “no movement to and from A positions” cover pronoun slots (IA-3 vs IA-1*)? Where are indirect-speech complements treated? Is FOC-LINK a parsing default (Ch. 1 (4b))? A visual tree check of Ch. 3–9; “complete functional complex” in the technical sense? (T35c done 2026-10-18, 8:46pm: subject on the head)Willie's framework §10; Young & Morgan 1987; Kari 1976; Jelinek 1990; Willie Ch. 3, 5, 7
T29a–j(new 2026-10-16, 12:46pm) After restating A/B as spell-out: is the verb word a spell-out cycle, and what makes it one? Does the object c-command the subject inside the word (decides Hale/CYC vs antisymmetry)? Is Qual above or below M/S (Hale vs Goad & Travis)? Are da’ and doo phrases, as CYC requires? Does Rice's scope order for disjuncts clash with CYC's height order? Can the oblique-object and deictic-subject exceptions be derived? Which stacked clause-final particle orders occur? (T29a–j)CYC (rule sets A/B), AS-1, AS-2, STRATA timingRice 2000 scope tests (not opened); Speas 1991b and Harley 2011 (not opened); co-occurrence data for clause-final particles; Hale 2003 n. 12–13 page images in better copies. OPEN
T28a–f(new 2026-10-16, 4:46am) After the site-by-site Condition C re-run: does the Navajo subject pronoun move (V-P) or stay in Spec,v (V-H)? Which site is each noun at, sentence by sentence, and what in (4.119) could put its clause higher? Are left-edge nouns first-merged or moved, and does Condition C read the lower copy? Where do postpositional phrases sit? Can a postposition's pronoun bind out of its PP (A10)? Do Baker's licensing conditions (21)–(22) carry over to Navajo? (T28a–f)A3′, D6, FB-5, LINK-IN, LINK-vP + V-P, STRATA orderingSpeas 1990 Ch. 4; minimal pairs (T25a); W&J 2000 pp. 270–278; Baker 1996 (not opened). OPEN
T27a–p(new 2026-10-15, 9:50pm) Framework audit follow-ups: open the unopened originals (Hale 1973, Barss et al., Young & Morgan, Rice 2000, Kari, Platero, Speas 1990 Ch. 4, Hale & Keyser) and classify them; restate our rule sets A/B, the inverse, the Condition C derivations (T27i done 2026-10-16: Condition C by site), nominal adjunction, wh/chains and relatives in VPISH/TP/CP terms (T27a–p on the framework audit)Rule sets A/B, inverse, A3/D6, nominal adjunction, wh, relativesSources not opened (see §3). OPEN
T11Is the Navajo fourth person the same device as Algonquian obviation or the Inuit “fourth person”? Does any inverse elsewhere give sourced evidence for VPISH raising?RANK-OBV / PERSON-4 / BOUND-3R (compared §5); D-BIND, D-WCO, D-AMB (§6)Partly answered 2026-10-09 (afternoon): no. These are three different jobs. Algonquian obviation corresponds to Navajo yi-/bi- (Willie 1991:96, 216; Bruening 2005:21). The Greenlandic 3r is bound by the matrix subject (Lybech 2020:115, 156). Bruening's Passamaquoddy inverse is A-movement with reconstruction, which gives T10a a three-way prediction. The binding diagnostics probably cannot be built in Navajo (W&J (71) + 4-NO-INV). See T11a–T11f.

Added 2026-10-15 (9:50pm): framework audit, T27a–p; correction: Willie & Jelinek 2000 pp. 264–278 were opened (the 9:20pm “abstract only” was wrong).

Added 2026-10-15 (8:46pm), revised 9:20pm: Jelinek & Baker attach. Jelinek 1984 tree (28) is pre-VPISH notation for Warlpiri and gives no verdict on A3 (the 8:46pm “supports A3” is withdrawn); Baker's later n. 16 leaves Navajo open with a PP/vP option against A3; Baker 1996 unopened. Rows T26a–g (T26g new: restate the PAH in VPISH/TP/CP structure).

Added 2026-10-14 (4:46am): open questions T20a–T20f on the ranking behind the inverse (bi- for equal-ranked animates; any inanimate inverse subject; the ‘upon’ element in two verb words; entailment vs capacity reading of Witherspoon's ‘let’; two moving inanimates; graded absurdity, which bears on T6) are listed on the who acts on whom page. T6 note: Witherspoon reports graded judgments, (8) “much more absurd” than (6), (6) “only slightly more absurd” than (4) (1977:65), and quotes Hale on “relative acceptability” (1973:306).

Added 2026-10-13: open questions T19a–T19h on wh-questions (Willie 1991 vs W&J 2000 on fronting with equal ranks; two argument question words; why ‘where’ + ‘who’ is out; slot entailments vs animacy rank; a possessor question inside a place phrase; misnumbered examples; questions about a 4th-person participant; Dogrib long-distance wh) are listed on the wh-questions page. T18d is partly answered there (§7). The multiple-wh test (§2) has its first sourced fact.

Added 2026-10-11 (evening): open questions T18a–T18h on nominal adjunction (the ‘shooting’ override with no visible reason; lone-noun subjects with intransitive vs transitive verbs; world knowledge vs grammar; nesting in the inverse; the “2d” gloss of (1); plain l vs ł in ‘ruin’; whether a noun can interrupt the MVS; one free pronoun per clause) are listed on the nominal adjunction page. T2 has a new datum there (§8, §10).

Added 2026-10-11 (afternoon): open questions T17a–T17i on pleonastic ho- and psych verbs (is ho- a role or a slot without one; why *‘I like the town’; where areal ho- sits in the template, a VPISH test; whether the Experiencer belongs to the verb's schema; inchoative psych verbs (Croft/Dowty); the (1f)/(41e)/(2e) marks; the glosses of béé and hózin; -’a animate vs inanimate; the agentive passive vs the inverse) are listed on the pleonastic & psych page. T2 has a new datum there (§9).

Added 2026-10-11 (early morning): open questions T16a–T16h on possessor raising (movement vs a new two-slot schema; Willie's “never 1st, 2nd, 4th” vs her own 1st-person (6c); whether a non-argument body part counts for animacy ranking; a swapped gloss pair (19)/(20); the (23) translation; (11.b′) without yi-; the incorporated noun shapes; Muskogean and other-Athabaskan comparisons, kept apart) are listed on the possessor raising page. T11d is partly answered there (§7).

Added 2026-10-10 (evening): open questions T15a–T15g on the uses of the fourth person (why léiʼ licenses a noun; why the reportative 4 skips the postposition; whether “4 on every verb” diagnoses the reportative; the optative + lágo counterexample; ‘think’ with non-optative complements; Uyechi 1990; the 1st-person ‘cry’ misprint) are listed on the fourth person uses page.

Added 2026-10-10 (afternoon): open questions T14a–T14f on relative and nominalized clauses (why ‘say’, ‘believe’ and adjunct clauses differ; internal/external as position or case; two misprinted glosses; the enclitic allomorphs; Platero's dialect reports; *(39) as processing vs grammar) are listed on the relatives page.

Added 2026-10-10: open questions T13a–T13f on reported speech and person focus (respect vs quotative jiní; the 2nd-person slot on nijiní; whether booyeéł is inverse; whether the 3/3 default is set by the construction; other perception verbs; where the complement of ‘say’ sits) are listed on the reported speech page.

Added 2026-10-09 (evening): open questions T12a–T12f on =go clause chains (the indices in (9); whether the patient condition names a role or a position; why a new agent comes only from the first clause; indefinite agents; where =go sits; other construction types) are listed on the =go chains page.

Added 2026-10-09 (afternoon): open questions T11a–T11f on the cross-linguistic comparison (Slave/Sarcee/Apache; bound-variable possessors; the sharpened scope test; possessor b-/y-; cross-clause reference; Yupik) are listed on the fourth person compared page.

Added 2026-10-09: open questions T10a–T10f on number and scope across the inverse (the scope judgments for (58)–(59); where da- sits and what Mode/Subject agrees with; dabi- and "maximally referential" subjects; da- with a singular agent in (68); paucal readings; da- with a 4th-person agent) are listed on the number & scope page.

Added 2026-10-08 (evening): open questions T8a–T8g on the fourth person (VPISH test; Chs. 5–6; W&J (75)–(76) strings; Y&M paradigm; multiple-wh; spellings; gradience) are listed on the fourth-person page.

Added 2026-10-08: open questions T5a–T5g on the inverse (which slot bi- indexes; grammar vs usage; the VPISH scope test; why themes never invert; fronting and -ísh; spellings; Hale 1973) are listed on the inverse page.

Added 2026-10-07 (evening): open questions T4a–T4f on which intransitives take a plain ł-causer are listed on the plain ł-causer page. T4c is the same question as K1/K4.

Added 2026-10-07: open questions K1–K6 on positional and stative causatives are listed on the keep vs bring about page.

2. The multiple-wh test OPEN (needs native intuitions)

Grant's test sentence: who bet who what at what casino. No Navajo version is given here. None was found in an opened source, and this site does not invent one. Below are schematic forms and what each theory predicts.

Schematic:  [wh₁(-sh)]  [wh₂]  [wh₃]  [wh₄ P]  [ROOT-verb word]
            better      bettee  stake   casino (postpositional)

3. Chee 2025 frame counts OPEN

Chee 2025 (J. Child Lang. 52(5):1010–1035; opened in a prior project run) and Chee & Yazzie describe the "phrasal verb" as a postposition plus a (minimally inflected) verb. About 17–19% of verb tokens in their child/caregiver sample occur with a postposition (prior-run note). Examples recorded there: shił nízhóní ‘I like it’; bikááʼ ádą́ní ‘table’. Both forms were not re-checked this run, so they appear only on this page.

Why it matters for the schema. HYPOTHESIS If the participant inside the postposition is a slot in Σ (lexically selected), then per-verb frame counts should be small and stable: each verb has a fixed set of postpositions. If postpositional participants are rule-set-B adjuncts, the frames should be open, with many postpositions per verb at low frequency each. Data needed: for each verb in Chee's (or Grant's) corpus, the number of distinct postposition frames and the token count of each.

4. Open questions from Grant's project notes (§9, verbatim)

These seven are copied verbatim from the project file Navajo_Verb_Structure_Exploratory_Notes §9. That file comes from Grant's Grok project and is not an authoritative source. The right column says where this site takes each one up.

#Question (verbatim)On this site
1What is the precise inventory and semantic contribution of each classificatory stem series in contemporary Navajo, and how stable are the Proto-Patient entailments across series?Not addressed in v0. It would populate EρP for classificatory stems.
2How exactly do the yi-/bi- (and related) inverse markers interact with specific template slots, and what are the full conditions on their appearance?Conditions C4–C9 (schema). The slot is position IV (object). Interaction with oblique objects (W&J (50)) is not yet modeled.
3Can controlled elicitation of periphrastic four-participant risk, attitude, transfer, and payment meanings be carried out systematically? What strategies do speakers prefer?Multiple-wh test; needs fieldwork.
4Is there evidence (synchronic or historical) that classificatory stem series result from incorporation or fusion of earlier phrase-level material, or is stem-level lexicalization the more accurate characterization?Not addressed. Relevant to whether rule set A ever takes rule-set-B input.
5How should the "verb-centric" packaging difference be formalized relative to local linearization and phase structure, rather than as a classic head-directionality parameter?Directly: IRP (iv) and (i′) vs Rice (iii) on the headedness page. If B2 is a phase/spell-out domain, (i′) becomes "linearize functional heads initial within the first spell-out domain of V". That is a phase-based restatement, HYPOTHESIS.
6What reconstruction, binding, or scope diagnostics are available inside the verbal complex itself, and what do they show about the relative height of participants encoded by prefixes versus the stem?T3 and T5 above. This is the single most important data gap for both phenomena.
7How do mode/aspect morphology and the classificatory stem series jointly determine telicity and measuring-out?Not addressed. It would bear on the "inc" (incremental theme) entailment.

5. Honest gaps in v0