claims#parole-masque-vecu-traction-not-substrate

"Parole-masque" judgments turn on the operation's traction on a pre-objective lived world (monde primordial), not ground/no-ground, human/machine, or body-as-substrate

ID: parole-masque-vecu-traction-not-substrate Title: "Parole-masque" judgments turn on the operation's traction on a pre-objective lived world (monde primordial), not ground/no-ground, human/machine, or body-as-substrate Status: candidate Confidence: speculative Claim type: interpretive / structural-parallel Created: 2026-06-19 Updated: 2026-06-19 Sources: merleau-ponty-2020-probleme-de-la-parole, chouraqui-2021-body-and-embodiment, chouraqui-2014-ambiguity-and-absolute Wiki homes: speaking-spoken-speech, self-falsification, motor-intentionality

Claim

In Le Problème de la parole (1953–54) Merleau-Ponty treats masque as the savoir-verbal-extérieur / pseudo-personnalité pole of speech's "double fonction du langage," and states the mask-function universally ("la parole est plutôt masque qu'expression," raw L2289; "aucune [parole] ne l'est tout à fait," L2403) — so the parole parlante / parlée difference is a gradient indexed to a vécu-bearing speaker ("± près du vécu," L2393), not a mask/face binary. The criterion separating the poles is the operation (conquête de nouvelles significations vs savoir verbal extérieur; the writer, not the perceiving body, is its paradigm, L2875) — but the operation is necessary-but-not-sufficient: full speech owes its reach not to décentration alone (which yields only "personne apparente," L2297) but to its traction on a pre-objective lived world ("notre rapport préobjectif avec le monde," L2297 / L2403). The synthetic, contestable bridge (which MP and Chouraqui do not make): for judging generated speech as "parole-masque," the decisive axis is therefore neither ground/no-ground (Nietzsche/Chouraqui) nor human/machine nor body-as-substrate, but whether the speech has traction on a pre-objective lived world (a monde primordial) — so embodiment matters as having a monde primordial for speech to bring to light, not as substrate. A disembodied LLM performs the décentration / accès à signification transindividuelle that produces only the savoir-verbal-extérieur PRODUCT, while lacking the rapport préobjectif that parole pleine requires: it occupies the mask-pole as product, not concept.

Evidence

merleau-ponty-2020-probleme-de-la-parole — verified raw anchors: L2289 (mask stated universally); L2393 (the gradient "± près du vécu"); L2441–2447 (mask = "disjonction du vécu et du langage," "la parole se vide et ne constitue plus qu'un masque"); L2451 ("conquête du monde primordial"), L2455 ("Parole comme corps ni première ni seconde"); L2297 + L2403 (the load-bearers: décentration alone yields only "personne apparente [les bons élèves… de Piaget]"; parole pleine "≠ seulement décentration… mais… notre rapport préobjectif avec le monde"); L2856–2867 (the "double fonction" table; "il ne peut être conquête qu'en étant aussi sédimentation"); L2875 (parole créatrice "à son comble" in the writer); L699 (the language-acquisition zero-point is the vécu-bearing child — "le vécu est l'être même," read in its child-egocentrism / vécu indivis context, not as a freestanding ontological equation).
chouraqui-2021-body-and-embodiment — the body as a self-falsifying FORCE, "constitutive and constituted at once… (falsely) regards itself as Körper… defined not as a subject but as a process" (ch. 6–7, raw L1224 / L1406 / L1410 / L1412), and power-as-vulnerability / reciprocity (ch. 9; see the live claims#embodiment-disproves-sovereignty) supply the non-substantial "lived side" the axis presupposes — so the axis is not a metaphysical ground.
The Evidence above anchors the descriptive axis only. No Evidence bullet grounds the worth-ranking "AI is deficient"; the anchors support the structural description, not the normative verdict, which is this candidate's deliberately-capped interpretive overreach (held at speculative — see Counterpressure). The AI application is a synthetic extension MP/Chouraqui do not make; it must not inherit the PbP anchors' support.

Counterpressure / Limits

False friend (the primary limit). MP's masque is constitutively subject-indexed: narcissistic pseudo-personnalité (L2289); the aphasic's savoir verbal extérieur — a langage quasi automatique that survives "quand [la vraie dénomination] est relativement conservé[e]" (L2472), i.e. a structure lost, not never-had — "un masque de la désintégration" (L4350); and the writer's surinvestissement of a vécu. A subjectless token-generator instances none of these: it resembles the savoir-verbal-extérieur product seen from outside but does not instance MP's privative, subject-indexed concept. Cross-tradition cousin (registral divergence), NOT identity; the AI bridge must not inherit the PbP anchors' support nor raise the masque motif's weight. Cf. the candidate precedent claims#circulus-vitiosus-two-receptions (one phrase, two operations that rhyme, not identity).
The Keller mirror. MP's nearest textual mask-case — Helen Keller (L2403) — is still a vécu-bearer: she "reçoit le langage plutôt qu'elle ne le recrée à partir de son silence" because she lacks sensory materials ("faute de matériaux") for a vécu she has. The LLM is Keller's mirror — all materials, no silence; she is all silence, no materials. Same surface (reception, not re-creation), opposite genesis. So Keller is an illustrative foil, not evidence that "no preobjective relation → reception," and the LLM is further off MP's axis than Keller, not a Keller-grade case.
Ungroundable verdict (recursion ceiling). By Chouraqui's own thesis (Being is self-falsification; "this ontology is nothing but a sedimentation… a sedimentative event," self-falsification) and the live claims#being-is-power-chouraqui-perceptual-faith-shared-structure (perceptual faith institutes reality while believing it merely recognizes it), the normative verdict "AI is deficient" institutes the ranking while believing it recognizes it — a Fürwahrhalten, not a finding. The framework can DESCRIBE the structural difference but cannot GROUND the verdict.
vécu-as-instituted-force (open). If Chouraqui's body-as-FORCE makes the lived side itself an instituted "half-thing" with no unproduced floor, the threshold softens toward a gradient; defense: even an instituted force is "acted upon as it acts" (woundable/reversible), surviving as a relational threshold-locus, not a self-grounding magnitude.
Continuity reading (defeated, recorded). Since the mask-law is universal and "± près du vécu" admits no pure pole (L2403 "aucune ne l'est tout à fait"), one could read AI as merely far along ONE dial — defeated by the relatum point (the dial presupposes a vécu; L699, L2297), but the magnitude-vs-relatum status of the vécu is the contestable hinge.
Embodied-robot on-axis (open). A woundable, reversible, sensorimotor agent with a genuine pre-objective world-relation could in principle be ON the axis; this is an empirical/structural question, not settled by MP.

Payoff

Dissolves three false binaries at once (mask/face, ground/no-ground, human/machine) and relocates the criterion to a single textually-anchored axis — the operation's traction on a pre-objective lived world — that (a) explains why a disembodied LLM is structurally the savoir-verbal-extérieur pole without substance-chauvinism, (b) predicts the embodied-robot and self-modifier cases as gradient/open rather than categorically excluded, and (c) is something a reader could not extract from the concept pages alone — it required reading PbP's masque occurrences against Chouraqui's force/sediment fold. Surfaced by a 5-hypothesis adversarial debate (this session).

Status History

2026-06-19 — created at candidate (5-hypothesis adversarial debate; unanimous RATIFY-WITH-FIX). 3-test gate met: contestable (the magnitude-vs-relatum hinge; the continuity reading is the live rival); evidence anchored at the verified PbP raw lines above; counterpressure recorded (false-friend, recursion ceiling, vécu-as-force, Keller-mirror). Held at candidate (not live) because the AI application is a synthetic extension MP/Chouraqui do not make, carrying an explicit false-friend caution. Durable gate-note: this candidate's AI / normative bridge can never reach supported — Test 5 fails by construction, since its strongest counter-position is the recursion ceiling that the framework's own ontology concedes. The TEXTUAL lemma (MP's langage-masque + the double-function, vécu-indexed gradient) is separately STRUCTURAL/live-grade and homed in motifs + speaking-spoken-speech, not in this candidate.