Hungarian Case Inventory @cite{kenesei-vago-fenyvesi-1998} @cite{rounds-2001} @cite{caha-2008}

Hungarian's case inventory per the two standard reference grammars: @cite{kenesei-vago-fenyvesi-1998} list 18 cases (see their Symbols table); @cite{rounds-2001} adds 4 less-productive cases (temporal -kor, distributive-temporal -nta, sociative -stul / -stül, locative fossilized -t / -tt) for a total of 22. All marking is via agglutinative suffixes.

Both reference grammars converge on three substantive points:

1. No morphological genitive. -nak / -nek is exclusively glossed as dative — even in possessive constructions where the possessor is "extracted" into a non-adjacent position. @cite{kenesei-vago-fenyvesi-1998} §1.10 explicitly attributes the analysis to Szabolcsi 1986/1992, 1994 and frames the possessor as the dative possessor (not GEN). @cite{caha-2008} §5 (pp. 266–267) likewise states verbatim: "Hungarian has nominative, accusative, dative, instrumental and a number of spatial cases, but no genitive ... possessor inside a Noun phrase ... is expressed as a dative, or nominative, depending on word-order, among other things." Possession is head-marked on the possessum; see Fragments/Hungarian/Possession.lean.

2. Local cases form a 3 × 3 matrix (interior / exterior / near × motion-toward / no-motion / motion-away) — see @cite{rounds-2001} §6.2's "Locative system: parameters of motion and space" table.

3. Hungarian is a known surface counterexample to Blake's hierarchy. @cite{caha-2008} fn. 8 cites Blake's own resolution: "the counterexamples are superficial, and are basically due to two factors: systematic syncretism (perhaps as in the case of Hungarian which uses dative to express possessor)..." Both Blake and Caha accept Hungarian as a typological exception explained by the dative-as-possessor analysis, not as a falsifying datum.

This Fragment exposes a 9-element Finset Core.Case capturing the broad case-functions that participate in Blake's hierarchy:

• Grammatical: NOM (∅), ACC (-t), DAT (-nak / -nek)
• Local — 3 × 3 matrix collapsed to direction only:
  • static (→ .loc): inessive (-ban / -ben), adessive (-nál / -nél), superessive (-n / -on / -en / -ön)
  • source (→ .abl): elative (-ból / -ből), ablative (-tól / -től), delative (-ról / -ről)
  • goal (→ .all): illative (-ba / -be), allative (-hoz / -hez / -höz), sublative (-ra / -re)
• Other: INST (-val / -vel), COM (= INS-form per @cite{kenesei-vago-fenyvesi-1998}; separate Finset element here), CAUS (-ért, "causal-final")

What Core.Case can express but this inventory omits:

• .Sup, .Sub, .Del (PascalCase UD spatial constructors) would preserve Hungarian's surface-row local cases distinctly rather than collapsing to the direction-only triple.
• .ess (essive-modal -ul / -ül), .transl (translative -vá / -vé), .ter (terminative -ig), .tem (temporal -kor) — all attested in both grammars, omitted here.
• ESS-FOR (-ként, "essive-formal", listed separately by both grammars) has no Core.Case constructor.
• DISTR (-nként), per @cite{rounds-2001} §6.4, has no Core.Case constructor — the only Hungarian case the substrate genuinely cannot express.

def Fragments.Hungarian.Case.caseInventory : Finset Core.Case

Hungarian case inventory: 9-element sample of Core.Case. The omission of .gen reflects the descriptive-grammar consensus (@cite{kenesei-vago-fenyvesi-1998}, @cite{rounds-2001}) and @cite{caha-2008} §5 — Hungarian has no morphological genitive.

Equations

• Fragments.Hungarian.Case.caseInventory = {UD.Case.nom, UD.Case.acc, UD.Case.dat, UD.Case.loc, UD.Case.abl, UD.Case.all, UD.Case.inst, UD.Case.com, UD.Case.caus}

theorem Fragments.Hungarian.Case.inventory_fails_strict : ¬Core.Case.IsValidInventory caseInventory

Hungarian fails Blake's strict contiguity at rank 5 (GEN), since the inventory has DAT (rank 4) without GEN. Parallels Finnish's failure at rank 4 (DAT) — Fragments.Finnish.Case.inventory_fails_strict. @cite{caha-2008} §5 (pp. 266–267) cites Hungarian as the textbook surface counterexample to Blake, resolved (per Blake fn. 8) by the dative-as-possessor syncretism.