Documentation

Linglib.Phenomena.Imperatives.Studies.RuytenbeekEtAl2017

@cite{ruytenbeek-etal-2017}: Indirect Request Processing, Sentence Types #

and Illocutionary Forces

Journal of Pragmatics 119 (2017) 46–62.

Two French eye-tracking experiments testing the literalist view that sentence types encode illocutionary force at the semantic level.

Core Finding #

Non-literalist theories are supported: directive illocutionary force is not encoded in sentence type but arises from shared semantic features between imperatives and deontic modals. Specifically:

Non-conventionalized indirect requests (Est-il possible de VP?) are processed as fast as conventionalized ones (Pouvez-vous VP?) and imperatives for directive interpretations.
Deontic necessity modals (Vous devez VP) receive directive force as readily as imperatives — same response times, no fixation on true/false buttons — unlike ability modals (Vous pouvez VP) or existential modals (Il est possible de VP).

Two Mechanisms for Directive Force #

The paper distinguishes two routes to directive illocutionary force:

Shared deontic semantics (@cite{kaufmann-2012}): Vous devez VP shares deontic necessity with imperatives → directive force is direct (Study 2). Modeled by directiveCompatible.
Convention of means (@cite{clark-1979}): Pouvez-vous VP? and Est-il possible de VP? question a preparatory condition for the request (addressee's ability) → directive force is indirect (Study 1). Not modeled by directiveCompatible — these constructions get directive force through pragmatic inference, not shared modal flavor.

Connection to Assert.lean #

The paper experimentally validates what Semantics.Modality.Assert already encodes: primaryFlavor .imperative = .deontic. @cite{kaufmann-2012}'s thesis — that imperatives have the semantics of deontic necessity modals — predicts that deontic necessity declaratives should receive directive force as readily as imperatives. Study 2 confirms exactly this.

Connection to SpeechActs.lean #

The paper reveals a gap in the SAPMood → IllocutionaryMood mapping: SAPMood.toIllocutionaryMood treats the mapping as 1-to-1, but indirect speech acts involve a mismatch — a declarative or interrogative sentence type receiving directive illocutionary force.

inductive Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.SentType :

Sentence types used in the experiments. These are morphosyntactic categories (sentence types), NOT illocutionary forces — the paper's core point is that these come apart.

imperative : SentType
canYouInterrog : SentType
possibleInterrog : SentType
ctrlInterrog : SentType
mustDeclarative : SentType
canDeclarative : SentType
possibleDecl : SentType
plainDeclarative : SentType

Instances For

instance Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.instDecidableEqSentType :

DecidableEq SentType

Equations

Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.instDecidableEqSentType x✝ y✝ = if h : x✝.ctorIdx = y✝.ctorIdx then isTrue ⋯ else isFalse ⋯

def Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.instBEqSentType.beq :

SentType → SentType → Bool

Equations

Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.instBEqSentType.beq x✝ y✝ = (x✝.ctorIdx == y✝.ctorIdx)

Instances For

instance Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.instBEqSentType :

Equations

Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.instBEqSentType = { beq := Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.instBEqSentType.beq }

instance Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.instReprSentType :

Equations

Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.instReprSentType = { reprPrec := Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.instReprSentType.repr }

def Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.instReprSentType.repr :

SentType → ℕ → Std.Format

Equations

One or more equations did not get rendered due to their size.

Instances For

def Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.SentType.mood :

SentType → Minimalism.Phenomena.SpeechActs.SAPMood

The morphosyntactic mood of each sentence type.

Equations

Instances For

def Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.SentType.modalFlavor :

SentType → Option Core.Modality.ModalFlavor

The contextually salient modal flavor contributed by the modal (if any).

NB: pouvoir and il est possible de are polysemous across epistemic, deontic, and circumstantial readings (see Fragments.French.Modals). The flavor assigned here is the contextually salient one in the experimental setting (2nd person, action-oriented context).

Equations

Instances For

def Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.SentType.modalForce :

SentType → Option Core.Modality.ModalForce

The modal force (if any).

Equations

Instances For

def Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.directiveCompatible (flavor : Core.Modality.ModalFlavor) :

A construction is directive-compatible when its modal flavor matches the imperative's primary flavor (deontic).

This models mechanism 1 (shared deontic semantics): deontic modals in declaratives receive directive force because they share the relevant semantic feature with imperatives (@cite{kaufmann-2012}).

This does NOT model mechanism 2 (convention of means / preparatory condition questioning): Pouvez-vous VP? and Est-il possible de VP? get directive force by questioning the addressee's ability, which is a preparatory condition for requests (@cite{clark-1979}). That mechanism operates via pragmatic inference, not modal flavor matching.

Equations

Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.directiveCompatible flavor = (flavor == Core.Modality.ModalFlavor.deontic)

Instances For

def Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.SentType.isDirectiveCompatible :

SentType → Bool

Directive compatibility for a sentence type via mechanism 1 (shared deontic semantics). Returns false for interrogative IRs, which use mechanism 2 (preparatory condition questioning) instead.

Equations

x✝.isDirectiveCompatible = match x✝.modalFlavor with | some f => Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.directiveCompatible f | none => false

Instances For

structure Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.CorpusDistribution :

Corpus usage distribution for a construction, classified by illocutionary force in context.

construction : String
nTokens : ℕ
directiveN : ℕ
questionN : ℕ
rhetoricalN : ℕ

Instances For

instance Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.instReprCorpusDistribution :

Repr CorpusDistribution

Equations

Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.instReprCorpusDistribution = { reprPrec := Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.instReprCorpusDistribution.repr }

def Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.instReprCorpusDistribution.repr :

CorpusDistribution → ℕ → Std.Format

Equations

One or more equations did not get rendered due to their size.

Instances For

def Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.CorpusDistribution.directiveRate (d : CorpusDistribution) :

Equations

d.directiveRate = ↑d.directiveN / ↑d.nTokens

Instances For

def Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.frantextCanYou :

CorpusDistribution

Frantext corpus data: Pouvez-vous VP? with singular addressee. N = 365. Directive uses = 71%, questions = 25%, rhetorical = 4%.

Equations

Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.frantextCanYou = { construction := "Pouvez-vous VP?", nTokens := 365, directiveN := 259, questionN := 91, rhetoricalN := 15 }

Instances For

def Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.frantextPossible :

CorpusDistribution

Frantext corpus data: Est-il possible de VP? with singular addressee. N = 63. Directive uses = 16%, questions = 70%, rhetorical = 14%.

Equations

Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.frantextPossible = { construction := "Est-il possible de VP?", nTokens := 63, directiveN := 10, questionN := 44, rhetoricalN := 9 }

Instances For

theorem Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.frantextCanYou_total :

frantextCanYou.directiveN + frantextCanYou.questionN + frantextCanYou.rhetoricalN = frantextCanYou.nTokens

Corpus category counts sum to the total.

theorem Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.frantextPossible_total :

frantextPossible.directiveN + frantextPossible.questionN + frantextPossible.rhetoricalN = frantextPossible.nTokens

theorem Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.canYou_more_conventionalized :

frantextCanYou.directiveRate > frantextPossible.directiveRate

Pouvez-vous VP? is significantly more conventionalized as a directive construction than Est-il possible de VP? (χ²(2, N=428) = 66.75, p < 0.001).

Study 1 (N = 41) tests whether non-conventionalized indirect requests can be indirect but primary. 24 trials: 6 imperatives, 6 control interrogatives, 6 Can you VP?, 6 Is it possible to VP?. Participants hear French sentences paired with a grid of colored shapes and respond by either moving a shape (directive) or clicking yes/no (question).

For *Can you* and *Is it possible* interrogatives, half the trials
have the correct answer = yes and the move is possible; for the other
half the correct answer = no. Analysis restricted to move-possible
trials: 6/2 × 41 = 123 per interrogative type.

Key finding: directive interpretations of both *Can you* and *Is it
possible* sentences are processed as fast as imperatives, with no
fixation on the yes/no area for either.

inductive Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.Study1Response :

Response types in Study 1.

move : Study1Response
yesNo : Study1Response

Instances For

instance Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.instDecidableEqStudy1Response :

DecidableEq Study1Response

Equations

Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.instDecidableEqStudy1Response x✝ y✝ = if h : x✝.ctorIdx = y✝.ctorIdx then isTrue ⋯ else isFalse ⋯

instance Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.instBEqStudy1Response :

BEq Study1Response

Equations

Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.instBEqStudy1Response = { beq := Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.instBEqStudy1Response.beq }

def Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.instBEqStudy1Response.beq :

Study1Response → Study1Response → Bool

Equations

Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.instBEqStudy1Response.beq x✝ y✝ = (x✝.ctorIdx == y✝.ctorIdx)

Instances For

def Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.instReprStudy1Response.repr :

Study1Response → ℕ → Std.Format

Equations

One or more equations did not get rendered due to their size.

Instances For

instance Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.instReprStudy1Response :

Repr Study1Response

Equations

Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.instReprStudy1Response = { reprPrec := Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.instReprStudy1Response.repr }

structure Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.Study1Result :

Study 1 result: mean response time (ms) and proportion of directive (move) responses for each sentence type.

RTs are β coefficients from the linear mixed effects model (exact from the paper's text). Proportions are approximate, estimated from Fig. 3 (exact counts not reported).

sentType : SentType
moveRT : ℕ
yesNoRT : ℕ
moveProp : ℚ

Instances For

instance Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.instReprStudy1Result :

Repr Study1Result

Equations

Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.instReprStudy1Result = { reprPrec := Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.instReprStudy1Result.repr }

def Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.instReprStudy1Result.repr :

Study1Result → ℕ → Std.Format

Equations

One or more equations did not get rendered due to their size.

Instances For

def Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.study1Imperative :

Equations

One or more equations did not get rendered due to their size.

Instances For

def Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.study1CanYou :

Equations

One or more equations did not get rendered due to their size.

Instances For

def Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.study1Possible :

Equations

One or more equations did not get rendered due to their size.

Instances For

def Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.study1CtrlInterrog :

Equations

One or more equations did not get rendered due to their size.

Instances For

theorem Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.study1_directive_rts_similar :

have imp := study1Imperative.moveRT; have can := study1CanYou.moveRT; have pos := study1Possible.moveRT; (↑can - ↑imp).natAbs < 200 ∧ (↑pos - ↑imp).natAbs < 200 ∧ (↑can - ↑pos).natAbs < 200

Study 1 key finding: no significant difference in RT between imperative, Can you, and Is it possible for directive (move) interpretations (all p's > 0.99 in post hoc comparisons).

theorem Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.study1_canYou_more_directive :

study1CanYou.moveProp > study1Possible.moveProp

Study 1: Can you elicits more directive interpretations than Is it possible (β = 0.79, z = 2.031, p = 0.043).

theorem Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.study1_no_rt_penalty :

study1Possible.moveRT ≤ study1CanYou.moveRT

Study 1: despite conventionalization difference, directive RTs are indistinguishable — non-conventionalized IRs don't require extra processing. This contradicts the literalist prediction that non-conventionalized IRs must activate the question force first.

theorem Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.study1_canYou_question_slower_than_ctrl :

study1CanYou.yesNoRT > study1CtrlInterrog.yesNoRT

Study 1: question interpretations of Can you take longer than control interrogatives (β = 4729 vs 3707, t(29.34) = 3.49, p = 0.03). The conventionalized directive reading BLOCKS the question reading.

The strongest anti-literalist evidence: fixation duration on the yes/no buttons (AOI). If the literalist view is correct, directive interpretations of interrogatives should FIRST activate the question force, yielding fixation on the yes/no area. The data show the opposite.

Fixation durations from Fig. 5 (approximate, in ms):
- Imperatives (move): ~5ms (near zero)
- *Can you* (move): ~5ms (near zero)
- *Is it possible* (move): ~0ms
- Control interrogatives (yes/no): ~280ms
- *Can you* (yes/no): ~280ms
- *Is it possible* (yes/no): ~230ms

The linear mixed effects model found no difference between control
interrogatives and question interpretations of *Can you* and *Is it
possible* (χ²(2) = 1.66, p = 0.43).

structure Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.FixationResult :

Fixation data for a sentence type and response type.

sentType : SentType
isDirective : Bool
fixationMs : ℕ

Instances For

instance Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.instReprFixationResult :

Repr FixationResult

Equations

Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.instReprFixationResult = { reprPrec := Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.instReprFixationResult.repr }

def Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.instReprFixationResult.repr :

FixationResult → ℕ → Std.Format

Equations

One or more equations did not get rendered due to their size.

Instances For

def Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.fix1_imp_move :

Equations

Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.fix1_imp_move = { sentType := Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.SentType.imperative, isDirective := true, fixationMs := 5 }

Instances For

def Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.fix1_can_move :

Equations

Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.fix1_can_move = { sentType := Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.SentType.canYouInterrog, isDirective := true, fixationMs := 5 }

Instances For

def Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.fix1_pos_move :

Equations

Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.fix1_pos_move = { sentType := Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.SentType.possibleInterrog, isDirective := true, fixationMs := 0 }

Instances For

def Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.fix1_can_yesno :

Equations

One or more equations did not get rendered due to their size.

Instances For

def Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.fix1_pos_yesno :

Equations

One or more equations did not get rendered due to their size.

Instances For

def Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.fix1_ctrl_yesno :

Equations

One or more equations did not get rendered due to their size.

Instances For

theorem Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.study1_directive_no_fixation :

fix1_imp_move.fixationMs < 10 ∧ fix1_can_move.fixationMs < 10 ∧ fix1_pos_move.fixationMs < 10

Directive interpretations show virtually NO fixation on the yes/no buttons — the question meaning is not activated.

theorem Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.study1_question_has_fixation :

fix1_can_yesno.fixationMs > 200 ∧ fix1_pos_yesno.fixationMs > 200 ∧ fix1_ctrl_yesno.fixationMs > 200

Question interpretations DO show fixation on yes/no buttons.

theorem Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.study1_fixation_gap :

fix1_can_yesno.fixationMs > fix1_can_move.fixationMs * 40 ∧ fix1_pos_yesno.fixationMs > fix1_pos_move.fixationMs * 40

The fixation gap is massive: question interpretations fixate on the response buttons 40–60× longer than directive interpretations. This is the smoking gun against literalism.

Study 2 (N = 40) tests whether deontic necessity modals (Vous devez VP) are processed as directives in the same way as imperatives. 24 trials: 3 You must, 3 control imperatives, 6 You can/may, 6 It is possible, 6 control declaratives. Task identical to Study 1, but yes/no replaced by true/false (enabling both directive and assertive responses).

Trial counts: Must and imperatives always allow move (3 × 40 = 120 each).
Can and Possible: half allow move (6/2 × 40 = 120 each).

Key finding: *You must* sentences receive overwhelmingly directive
interpretations, just like imperatives, with identical response times.
The paper reports n = 21 true/false responses to Must (out of 120).

inductive Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.Study2Response :

Response types in Study 2.

move : Study2Response
trueFalse : Study2Response

Instances For

instance Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.instDecidableEqStudy2Response :

DecidableEq Study2Response

Equations

Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.instDecidableEqStudy2Response x✝ y✝ = if h : x✝.ctorIdx = y✝.ctorIdx then isTrue ⋯ else isFalse ⋯

instance Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.instBEqStudy2Response :

BEq Study2Response

Equations

Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.instBEqStudy2Response = { beq := Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.instBEqStudy2Response.beq }

def Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.instBEqStudy2Response.beq :

Study2Response → Study2Response → Bool

Equations

Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.instBEqStudy2Response.beq x✝ y✝ = (x✝.ctorIdx == y✝.ctorIdx)

Instances For

def Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.instReprStudy2Response.repr :

Study2Response → ℕ → Std.Format

Equations

One or more equations did not get rendered due to their size.

Instances For

instance Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.instReprStudy2Response :

Repr Study2Response

Equations

Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.instReprStudy2Response = { reprPrec := Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.instReprStudy2Response.repr }

structure Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.Study2Result :

Study 2 result for each sentence type.

RTs are β coefficients from the linear mixed effects model (exact from text). Proportions use 120 as denominator (3 × 40 for must/imperative, 6/2 × 40 for can/possible). Numerators for must are derived from the text (n = 21 true/false out of 120 → 99 move). Numerators for can and possible are approximate from Fig. 6.

sentType : SentType
moveRT : ℕ
trueFalseRT : ℕ
moveProp : ℚ

Instances For

instance Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.instReprStudy2Result :

Repr Study2Result

Equations

Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.instReprStudy2Result = { reprPrec := Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.instReprStudy2Result.repr }

def Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.instReprStudy2Result.repr :

Study2Result → ℕ → Std.Format

Equations

One or more equations did not get rendered due to their size.

Instances For

def Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.study2Imperative :

Equations

One or more equations did not get rendered due to their size.

Instances For

def Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.study2Must :

Equations

One or more equations did not get rendered due to their size.

Instances For

def Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.study2Can :

Equations

One or more equations did not get rendered due to their size.

Instances For

def Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.study2PossibleDecl :

Equations

One or more equations did not get rendered due to their size.

Instances For

theorem Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.study2_must_equals_imperative_rt :

(↑study2Must.moveRT - ↑study2Imperative.moveRT).natAbs < 200

Study 2 core finding: You must response times are indistinguishable from imperatives (p > 0.99 in post hoc comparison).

theorem Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.study2_must_mostly_directive :

study2Must.moveProp > 4 / 5

Study 2: You must receives the vast majority of directive interpretations.

theorem Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.study2_must_more_directive_than_can :

study2Must.moveProp > study2Can.moveProp

Study 2: You can and It is possible receive far fewer directive interpretations than You must.

theorem Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.study2_must_more_directive_than_possible :

study2Must.moveProp > study2PossibleDecl.moveProp

theorem Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.study2_can_more_directive_than_possible :

study2Can.moveProp > study2PossibleDecl.moveProp

Study 2: You can triggers more directive readings than It is possible (z = −3.29, p = 0.0028). Permission (deontic possibility) is closer to directive force than pure circumstantial possibility.

Fixation on the true/false buttons (Fig. 8, approximate in ms): - Imperatives (move): ~10ms - You must (move): ~15ms - You can (move): ~0ms - It is possible (move): ~0ms - You can (true/false): ~270ms - It is possible (true/false): ~265ms - Control declaratives (true/false, yes): ~270ms

Critically: imperatives, *You must* directive, and *You can*/*It is
possible* directive interpretations all show ~0ms fixation on the
true/false buttons. The assertive meaning is NOT activated.

def Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.fix2_imp_move :

Equations

Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.fix2_imp_move = { sentType := Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.SentType.imperative, isDirective := true, fixationMs := 10 }

Instances For

def Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.fix2_must_move :

Equations

One or more equations did not get rendered due to their size.

Instances For

def Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.fix2_can_move :

Equations

Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.fix2_can_move = { sentType := Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.SentType.canDeclarative, isDirective := true, fixationMs := 0 }

Instances For

def Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.fix2_can_tf :

Equations

Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.fix2_can_tf = { sentType := Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.SentType.canDeclarative, isDirective := false, fixationMs := 270 }

Instances For

def Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.fix2_pos_tf :

Equations

Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.fix2_pos_tf = { sentType := Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.SentType.possibleDecl, isDirective := false, fixationMs := 265 }

Instances For

theorem Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.study2_directive_no_fixation :

fix2_imp_move.fixationMs < 20 ∧ fix2_must_move.fixationMs < 20 ∧ fix2_can_move.fixationMs < 20

Study 2: directive interpretations show virtually no fixation on the true/false buttons — the assertive/statement meaning is not activated. This holds for You must (a declarative!) just as for imperatives.

theorem Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.study2_statement_has_fixation :

fix2_can_tf.fixationMs > 200 ∧ fix2_pos_tf.fixationMs > 200

Study 2: statement interpretations DO show fixation on true/false.

theorem Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.study2_must_declarative_no_statement_activation :

SentType.mustDeclarative.mood = Minimalism.Phenomena.SpeechActs.SAPMood.declarative ∧ fix2_must_move.fixationMs < 20

The core anti-literalist evidence from Study 2: You must is a DECLARATIVE sentence, yet directive interpretations show the same near- zero fixation on the statement-response buttons as IMPERATIVES. If literalism were correct, interpreting a declarative as a directive should first activate the assertive force → fixation on true/false. It doesn't.

The paper's core theoretical contribution connects to Assert.lean: imperatives and deontic necessity modals share a modal flavor, which is why they both license directive force (mechanism 1).

theorem Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.imperative_must_share_flavor :

SentType.imperative.modalFlavor = SentType.mustDeclarative.modalFlavor

Imperatives and deontic necessity modals share the same modal flavor. This is the semantic basis for the Study 2 finding that You must VP is processed identically to imperatives.

@cite{kaufmann-2012} makes this claim explicitly; the paper validates it experimentally.

theorem Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.imperative_must_share_force :

SentType.imperative.modalForce = SentType.mustDeclarative.modalForce

Imperatives and deontic necessity modals share the same modal force.

theorem Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.assert_imperative_is_deontic :

Semantics.Modality.Assert.primaryFlavor Core.Discourse.IllocutionaryMood.imperative = Core.Modality.ModalFlavor.deontic

The primaryFlavor from Assert.lean already encodes the imperative–deontic link: imperative speech acts have deontic content. This is the theory-layer fact that the paper experimentally validates.

theorem Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.imperative_directive_compatible :

SentType.imperative.isDirectiveCompatible = true

Both imperatives and You must declaratives are directive-compatible (mechanism 1). Follows from both having deontic modal flavor.

theorem Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.must_directive_compatible :

SentType.mustDeclarative.isDirectiveCompatible = true

theorem Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.can_not_directive_compatible :

SentType.canDeclarative.isDirectiveCompatible = false

Ability and existential possibility modals are NOT directive-compatible via mechanism 1. They have circumstantial (not deontic) flavor. This predicts the Study 2 finding: You can and It is possible receive fewer directive interpretations than You must.

NB: these constructions CAN still receive directive force via mechanism 2 (questioning a preparatory condition), which explains the non-zero directive rate (~30% for can, ~13% for possible).

theorem Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.possible_not_directive_compatible :

SentType.possibleDecl.isDirectiveCompatible = false

Directive compatibility (mechanism 1) correctly predicts the ranking of directive response rates in Study 2: deontic-flavored constructions (imperative, you must) get high directive rates; circumstantial- flavored constructions (you can, it is possible) get low rates.

The paper demonstrates that SAPMood.toIllocutionaryMood is the DEFAULT force mapping, not the only possible one. Indirect speech acts involve a sentence type receiving a non-default illocutionary force.

theorem Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.declarative_default_not_directive :

Minimalism.Phenomena.SpeechActs.SAPMood.declarative.toIllocutionaryMood ≠ Core.Discourse.IllocutionaryMood.imperative

The default illocutionary force for declaratives is NOT directive.

theorem Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.must_declarative_force_mismatch :

SentType.mustDeclarative.mood = Minimalism.Phenomena.SpeechActs.SAPMood.declarative ∧ SentType.mustDeclarative.isDirectiveCompatible = true

But deontic necessity declaratives CAN receive directive force. The force mismatch (declarative sentence type + directive force) is exactly what the paper demonstrates for Vous devez VP.

This is modeled by the isDirectiveCompatible predicate, which bypasses sentence type and checks the modal semantics directly.

theorem Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.canYou_force_mismatch :

SentType.canYouInterrog.mood = Minimalism.Phenomena.SpeechActs.SAPMood.interrogative ∧ study1CanYou.moveProp > 1 / 2

Interrogative IRs also exhibit force mismatch: interrogative sentence type with directive force (via mechanism 2).

The study's SentType.modalFlavor assignments are not stipulated in isolation — they derive from the fragment entries in Fragments.French.Modals. These bridge theorems ensure the study and fragment stay in sync: changing a fragment entry's flavor list will break the corresponding theorem here.

theorem Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.must_flavor_from_fragment :

Core.Modality.ModalFlavor.deontic ∈ Fragments.French.Modals.devoir.flavors ∧ SentType.mustDeclarative.modalFlavor = some Core.Modality.ModalFlavor.deontic

Vous devez VP uses devoir, which has deontic as an available flavor. The study assigns deontic to .mustDeclarative.

theorem Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.can_flavor_from_fragment :

Core.Modality.ModalFlavor.circumstantial ∈ Fragments.French.Modals.pouvoir.flavors ∧ SentType.canDeclarative.modalFlavor = some Core.Modality.ModalFlavor.circumstantial

Vous pouvez VP uses pouvoir. The study assigns circumstantial to .canDeclarative (the default non-epistemic reading in 2nd person declaratives — ability/permission). Pouvoir has circumstantial in its flavor inventory.

theorem Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.possible_flavor_from_fragment :

Core.Modality.ModalFlavor.circumstantial ∈ Fragments.French.Modals.ilEstPossible.flavors ∧ SentType.possibleDecl.modalFlavor = some Core.Modality.ModalFlavor.circumstantial

Il est possible de VP uses the impersonal construction. The study assigns circumstantial to .possibleDecl.

theorem Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.fragment_force_explains_study2 :

Fragments.French.Modals.devoir.force = Core.Modality.ModalForce.necessity ∧ Fragments.French.Modals.pouvoir.force = Core.Modality.ModalForce.possibility ∧ study2Must.moveProp > study2Can.moveProp

The force contrast between devoir and pouvoir is the structural explanation for the Study 2 asymmetry: necessity (obligation) licenses directive force more readily than possibility (permission).

theorem Phenomena.Imperatives.Studies.RuytenbeekEtAl2017.devoir_matches_imperative_speech_act :

{ force := Core.Modality.ModalForce.necessity, flavor := Core.Modality.ModalFlavor.deontic } ∈ Fragments.French.Modals.devoir.forceFlavors ∧ Semantics.Modality.Assert.primaryFlavor Core.Discourse.IllocutionaryMood.imperative = Core.Modality.ModalFlavor.deontic

The fragment's force-flavor pair for devoir includes deontic necessity — the same combination as the imperative speech act.