Muñoz Pérez (2026) — Stylistic Applicatives in Chilean Spanish #
Grammaticality judgments from [MP26] "Stylistic applicatives: A lens into the nature of anticausative SE" (Glossa 11(1)).
Main declarations #
Judgment,CliticPattern,DativeCliticPerson— empirical data typesspanishFissionRule— instantiation of the genericMorphology.DM.FissionRulewith Chilean-Spanish-specific datavoice_semantically_vacuous— re-export ofMinimalist.nonThematic_no_semanticsthree_way_synonymy_from_vacuity,fission_person_restriction,stylLE_requires_inchoative,unmarked_blocks_stylLE— bridge theorems for the empirical properties of stylistic LE the Fission analysis accounts for
Implementation notes #
Acceptability follows the project-canonical Features.Acceptability
six-level taxonomy. Paper-internal * maps to .unacceptable and the
unmarked judgment maps to .ok.
Key data points #
Three-way synonymy (exx. 7–12): For marked anticausatives with 1SG/2SG dative, three clitic patterns are interchangeable:
- SE + CL_dat: se me rompió "it broke on me"
- CL_dat + LE: me le rompió
- SE + CL_dat + LE: se me le rompió
Person restriction (exx. 15–19, cerrar la ventana): Stylistic LE is available only with 1SG (me) and 2SG (te), not 3SG (le), 1PL (nos), 2/3PL (les).
Marking restriction (exx. 39–44): Stylistic LE requires SE-marked (or optionally SE-marked) anticausatives. Unmarked anticausatives (mejorar) block it.
Negative controls (exx. 13b, 14b): quejarse and impersonal SE reject the me le pattern; the stylistic le is not a free dative.
Data types #
A clitic pattern in an anticausative construction.
- se_cl : CliticPattern
SE + dative clitic: se me rompió.
- cl_le : CliticPattern
Dative clitic + LE: me le rompió (stylistic applicative).
- se_cl_le : CliticPattern
SE + dative clitic + LE: se me le rompió.
Instances For
Equations
- MunozPerez2026.instDecidableEqCliticPattern x✝ y✝ = if h : x✝.ctorIdx = y✝.ctorIdx then isTrue ⋯ else isFalse ⋯
Equations
Equations
- One or more equations did not get rendered due to their size.
Instances For
Person of the dative clitic.
- first_sg : DativeCliticPerson
me
- second_sg : DativeCliticPerson
te
- third_sg : DativeCliticPerson
le
- first_pl : DativeCliticPerson
nos
- third_pl : DativeCliticPerson
les
Instances For
Equations
- MunozPerez2026.instDecidableEqDativeCliticPerson x✝ y✝ = if h : x✝.ctorIdx = y✝.ctorIdx then isTrue ⋯ else isFalse ⋯
Equations
- One or more equations did not get rendered due to their size.
Instances For
Equations
A single grammaticality judgment from the paper.
- exNumber : String
Example number in the paper.
- verb : String
The verb in citation form.
- pattern : CliticPattern
The clitic pattern.
- dativePerson : DativeCliticPerson
Person of the dative clitic.
- acceptability : Features.Acceptability
Acceptability per
Features.Acceptability.
Instances For
Equations
- One or more equations did not get rendered due to their size.
Instances For
Equations
- MunozPerez2026.instReprJudgment = { reprPrec := MunozPerez2026.instReprJudgment.repr }
Equations
- One or more equations did not get rendered due to their size.
- MunozPerez2026.instBEqJudgment.beq x✝¹ x✝ = false
Instances For
Equations
Three-way synonymy data (exx. 7–12) #
romper "break" with 1SG dative: all three patterns OK.
Equations
- One or more equations did not get rendered due to their size.
Instances For
Equations
- One or more equations did not get rendered due to their size.
Instances For
Equations
- One or more equations did not get rendered due to their size.
Instances For
hundir "sink" with 1SG dative.
Equations
- One or more equations did not get rendered due to their size.
Instances For
Equations
- One or more equations did not get rendered due to their size.
Instances For
caer "fall" with 1SG dative.
Equations
- One or more equations did not get rendered due to their size.
Instances For
Equations
- One or more equations did not get rendered due to their size.
Instances For
Equations
- One or more equations did not get rendered due to their size.
Instances For
morir "die" with 1SG dative.
Equations
- One or more equations did not get rendered due to their size.
Instances For
Equations
- One or more equations did not get rendered due to their size.
Instances For
Equations
- One or more equations did not get rendered due to their size.
Instances For
Negative controls (exx. 13b, 14b) #
Crucially, the me le pattern is NOT freely available — it is rejected with the inherently reflexive verb quejarse "complain" (ex. 13b) and with impersonal SE plus an argumental dative (ex. 14b). These witnesses keep the dataset honest: stylistic LE depends on the marked-anticausative structure, not on phonological adjacency.
quejarse "complain" rejects the me le pattern (ex. 13b).
Equations
- One or more equations did not get rendered due to their size.
Instances For
Impersonal SE + argumental dative rejects the me le pattern (ex. 14b).
Equations
- One or more equations did not get rendered due to their size.
Instances For
Negative-control judgments.
Equations
Instances For
Person restriction data (exx. 15–19, cerrar la ventana) #
1SG: stylistic LE is OK (ex. 15b Me le cerró la ventana).
Equations
- One or more equations did not get rendered due to their size.
Instances For
2SG: stylistic LE is OK (ex. 16b Te le cerró la ventana).
Equations
- One or more equations did not get rendered due to their size.
Instances For
3SG: stylistic LE is BLOCKED (ex. 17b Le le cerró la ventana).
Equations
- One or more equations did not get rendered due to their size.
Instances For
1PL: stylistic LE is BLOCKED (ex. 18b Nos le cerró la ventana).
Equations
- One or more equations did not get rendered due to their size.
Instances For
2/3PL: stylistic LE is BLOCKED (ex. 19b Les le cerró la ventana).
Equations
- One or more equations did not get rendered due to their size.
Instances For
Person restriction data collected.
Equations
Instances For
Marking restriction data (exx. 39–44) #
quebrar (marked SE) licenses stylistic LE (ex. 39b Me le quebró el florero).
Equations
- One or more equations did not get rendered due to their size.
Instances For
mejorar (unmarked) does NOT license stylistic LE (ex. 40b *Me le mejoró el sueldo).
Equations
- One or more equations did not get rendered due to their size.
Instances For
hervir (optional SE) DOES license stylistic LE (ex. 44a Me le hirvió el agua).
Equations
- One or more equations did not get rendered due to their size.
Instances For
Data verification #
All three-way synonymy patterns are grammatical for 1SG.
Person restriction: exactly 1SG and 2SG are grammatical.
Person restriction: exactly 3SG, 1PL, 3PL are ungrammatical.
The person paradigm uses cerrar (ex. 15), not caer (ex. 9).
Marking restriction: marked/optional → OK, unmarked → blocked.
Negative controls are present and uniformly unacceptable. Drift sentry:
if the me le pattern were ever miscoded as .ok, this fails.
Spanish Fission instantiation #
The Spanish-specific realization output of Fission: two clitic
positions. Cl₁ surfaces person features (me/te), Cl₂ is
invariably le.
- cl1Form : String
Cl₁: bears person features.
- cl2Form : String
Cl₂: bears case features.
Instances For
Equations
Equations
- One or more equations did not get rendered due to their size.
Instances For
Equations
- One or more equations did not get rendered due to their size.
Instances For
The stylistic applicative Fission rule for Chilean Spanish ([MP26] rule 55).
Instantiates the generic Morphology.DM.FissionRule with
Spanish-specific data:
- Context: inchoative verbal-head sequence (vGO ⌒ vBE)
- Bundle: [+PART, +SING] person (1SG or 2SG)
- Realization: Cl₁ = me/te (from [±AUTHOR]), Cl₂ = le (invariable)
Equations
- One or more equations did not get rendered due to their size.
Instances For
[MP26]'s PF condition (rule 58): the non-thematic
VoiceP projection must be overtly marked on the verb by a
reflexive clitic. se is the directly-merged reflexive; me/te
count because they are DAT-REFL syncretic (the paper's table 59)
and syncretic elements are indistinguishable for PF purposes.
(1PL nos is also syncretic but is filtered upstream: Fission
requires [+SING], so nos le is never generated.)
Equations
- MunozPerez2026.AnticausativePF out = (out.cl1Form = "me" ∨ out.cl1Form = "te" ∨ out.cl1Form = "se")
Instances For
Apply the Spanish stylistic applicative Fission rule.
Equations
- MunozPerez2026.applySpanishFission p heads = MunozPerez2026.spanishFissionRule.apply p heads
Instances For
Person restriction (paper §3.1) #
Fission applies only to 1SG and 2SG. DERIVED from [+PARTICIPANT, +SINGULAR] feature condition.
The person restriction matches the empirical data: Fission applies ↔ stylistic LE is grammatical.
Inchoative requirement (the context of rule 55) #
Stylistic LE requires inchoative context (vGO ∧ vBE). DERIVED from Fission's structural context condition.
Every Muñoz-Pérez verb that licenses stylistic LE has inchoative structure. DERIVED from the verb fragment.
Marking restriction #
Unmarked anticausatives block stylistic LE. DERIVED from the verb fragment: mejorar is unmarked and blocks LE.
Marked anticausatives license stylistic LE.
Optional SE-marking also licenses stylistic LE.
All Muñoz-Pérez verbs blocking stylistic LE are unmarked. DERIVED from the fragment data.
SE-optionality (the PF condition, rule 58) #
When Fission applies, the output clitic satisfies the PF marking condition (syncretic with reflexive), making SE optional.
The DAT-REFL syncretism that enables SE-optionality is present for exactly the persons where Fission applies.
Three-way synonymy #
Re-export of Minimalist.nonThematic_no_semantics in the Muñoz-Pérez
frame. SE is purely a PF marker — its presence or absence is
phonological, not semantic.
The empirical three-way synonymy is consistent with Voice
vacuity: the three .ok judgments co-hold with the proof that
Voice has no semantics (the judgments are data, not derived).
Fission verification #
Fission applies to 1SG in inchoative context.
Fission applies to 2SG in inchoative context.
Fission does NOT apply to 3SG (not [+PART]).
Fission does NOT apply in non-inchoative context (activity).
Fission does NOT apply in causative context (has vDO).
1SG Cl₁ is "me" (reflects [+AUTHOR]).
2SG Cl₁ is "te" (reflects [-AUTHOR]).
Cl₂ is always invariable "le".
Refutation of [KG09] #
K-G's reflexivization analysis predicts that every alternating verb
has SE in its anticausative form (cumulation of A and P spelled out as
SE). The verb-level predicate KoontzGarboden2009.kgPredictsSEMarked
formalises this chain. mejorar "improve" alternates while remaining
unmarked, falsifying the prediction. The paper's footnote 7 notes that
[KG09]'s own implementation restricts reflexivization
to SE-marked anticausatives, so the mejorar argument bites most
directly against the null-reflexive extension ([Chi04]).
Refutation of [KG09]: mejorar alternates but
is unmarked (no SE), against K-G's prediction that reflexivization
requires SE-spell-out. Closes the bridge from K-G's
reflexivization.involvesCumulation = true to a falsifying
Spanish verb.
Cross-framework comparisons #
The paper draws a second comparative argument — narrower than [MSK25]'s two-flavor Voice — that is not yet stated as a Lean theorem.
Todo #
MSK comparison as a real bridge theorem. Analogously,
MartinSchaeferKastner2025.seVoiceOptions : List VoiceFlavoris a list literal[.nonThematic, .reflexive]. A subset claim against any hand-written Muñoz list of flavors is decided bydecideover list literals — no real Voice-flavor mechanism is engaged. A genuine comparison requires deriving each paper's predicted flavor set from its analytical commitments (Voice-flavor licensing rules in MSK; Fission + Voice-vacuity in MunozPerez), then proving inclusion of the derived sets.vGO ⌒ vBE adjacency in
isInchoative(Phase D — substrate). Muñoz Pérez's Fission rule has the explicit context/vGO __ vBE, butSyntax/Minimalist/VerbalDecomposition.lean'sisInchoativechecks only set-membership (heads.contains .vGO), not adjacency. Deferred because the refactor touches 8 downstream consumer files. A focused session should add anapplPosfield (or similar adjacency witness) to the decomposition and audit each consumer site.Derive
licensesStylLEfrom structure (Phase D —Fragments/Spanish/Predicates.lean). The Fragment currently stipulateslicensesStylLE : Boolper verb; per CLAUDE.md's "derive, don't stipulate" rule it should be computed from existing structural fields, plausiblyisInchoative v.verbHead && v.anticausativeMarking ∈ [.marked, .optional]. Deferred pending Phase D'sisInchoativerefactor (the derivation needs adjacency-aware inchoativity to be empirically tight).Newman 2024 Feature Failure (Phase D — Minimalist substrate). Paper rule 60 grounds why the stylistic le's unchecked features do not crash the derivation. The Minimalist substrate currently has no
Derivationorcrashespredicates to formalise Feature Failure against; deferred pending those primitives.