Pre-registered, frozen measurement protocol testing whether the Qurʾanic corpus contains passages that instantiate the material and relational architecture of Sappho Fragment 31 and its pre-Qurʾanic reception history, at a rate and specificity comparable to independently recognized receptions. The protocol freezes a two-stage instrument before any Qurʾanic discovery scan: a coarse blind screen (nine material channel families M1–M9 from #1486; five relational clauses R1–R5 drafted without access to #1494) for corpus-scale retrieval and hypothesis-blind scoring, and the authentic twelve-clause operator of #1494 — recovered from the archive before this freeze — for confirmatory adjudication. Includes a contamination firewall over five previously seen verses, a corpus lock against the archive's own sovereign Tanzil seat (EA-CORPORA-02/08) with recorded SHA-256, positive and negative control panels scored before the Qurʾan is opened, Qurʾan-internal background sampling, blinding procedure, result classes, and falsification conditions. Tests P_material only; P_historical is never silently inferred. Infrastructure for future work in the Longinus/Philo/Josephus series.
Protocol record and PDF. The locked corpus resides at data/corpora/quran/original/quran-uthmani-tanzil-1.1.txt (verbatim-seat, EA-CORPORA-02/08).
Freeze date: 2026-08-22. Status: FROZEN at deposit; the Qurʾanic discovery scan has not begun and must not begin except under this protocol or a successor bearing its own version number. Target claim: material presence of the Fragment-31 transform-family in the Qurʾanic corpus. Not assumed: direct borrowing, conscious imitation, or a particular transmission route.
Two disclosures precede the design, recorded so that they cannot be papered over later.
First, the contamination event. A prior exploratory pass over the Qurʾan, made in working exchange before this protocol existed, surfaced Q 5:83, 7:143, 8:2, 39:23, and 41:20 as impressionistically promising. Those five verses therefore cannot count as discoveries under any outcome. They form a contaminated positive-interest set, quarantined by the firewall in §5 and usable only as a separately labeled comparison panel after the blind run.
Second, the #1494 recovery event. The drafting exchange could not recover the standalone body of deposit #1494 (The Transform Signature, Frozen) and deliberately declined to reconstruct it — a counterfeit operator silently standing in for the authentic one being the failure this program names. The relational clauses R1–R5 in §3 were accordingly drafted from downstream references only, and they are retained here verbatim as drafted: their independence from #1494 is now itself a documented property. Before this freeze, in the depositing session, the authentic #1494 body was recovered whole from the archive (AXN:0609.DATASET.☀️🌔🔴🌙🗂️🕑, body_status full). Its twelve clauses C1–C12 are therefore incorporated by reference as the confirmatory-stage operator (§9), and the drafted safeguard — that recovery of #1494 must not silently alter the screening instrument — is honored by freezing both layers side by side. Comparison summary, recorded at freeze: R1–R5 is a coarse projection of the C-operator's relational skeleton (R1≈C1 trigger-frame; R2–R3≈C2/C6/C9 conversion and displacement; R4≈C3/C11 threshold and cycle; R5≈C6/C12 consequence and compression); no contradiction between the layers was found; the C-operator is strictly finer (carrier-diagnostic shedding C4, surface-naming color C5, reactivation-by-another C7, agentless joints C8, second-order intensification C10).
Does the Qurʾanic corpus contain passages that instantiate the material and relational architecture frozen from Sappho 31 and its pre-Qurʾanic reception history at a rate and specificity comparable to independently recognized receptions of Fragment 31?
This formulation distinguishes two propositions. P_material: the operator is materially instantiated. P_historical: the Qurʾanic instantiation historically descends from the Sapphic reception tradition. QFO-01 tests P_material. A successful result licenses investigation of P_historical; P_historical is not silently inferred from it.
The nine channel families come directly from the blind-reconstruction protocol (#1486), registered before corpus contact with Josephus:
M1 JOLT — abrupt internal seizure, shock, heart-event, startle, immediate affective/autonomic change. M2 SIGHT — visual trigger, visual failure, blinding, altered seeing, sight made causally operative. M3 VOICE/TONGUE — speech failure, muteness, tongue dysfunction, compelled/restored speech, voice displaced. M4 FIRE/HEAT IN BODY — burning, heat, fire, inflammation, comparable embodied thermal alteration. M5 HEARING — roar, altered hearing, deafness, hearing made causally operative or occluded. M6 SWEAT/FLUID AUTONOMIC RESPONSE — sweating or close autonomic equivalent. M7 TREMOR — trembling, shaking, shivering, bodily quaking. M8 COLOR — pallor, greenness, flushing, blanching, explicit bodily colour-change. M9 DEATH/COLLAPSE THRESHOLD — near-death, faint, fall, unconsciousness, corpse-like state, explicit death-adjacency.
These are not a bag of keywords. Josephus demonstrated that a prose carrier can distribute the stations across clauses and persons, defeating section-level lexical co-occurrence while preserving the machine (#1493).
Scored independently of M1–M9. Drafted without access to #1494 (§0); retained verbatim.
R1 ENCOUNTER/RECEPTION — a subject encounters a person, voice, utterance, sign, appearance, or other perceivable object. R2 CAUSAL SOMATIC CONVERSION — the encounter materially alters the receiver; bodily change is caused by reception, not merely mentioned nearby. R3 CHANNEL DISPLACEMENT — at least one ordinary human channel (seeing, hearing, speaking, bodily composure) is occluded, seized, redirected, exchanged, or unusually activated. R4 THRESHOLD/ALTERED AGENCY — the event moves the subject toward or through loss of ordinary agency: collapse, death-adjacency, overpowering, possession, incapacitation — or a polarity-reversed restoration from such a state. R5 TRANSMISSION CONSEQUENCE — the transformed body/state becomes consequential for subsequent speech, witness, reception, remembrance, inscription, command, guidance, or another transmissible act.
Grounding: in Fragment 31 the bodily sequence culminates in death-adjacency; in Addition D and AJ 11 the same machine becomes cyclical — collapse, another's intervention, reopened speech, relapse (#1493); in BJ 1 polarity reverses while the relations remain operative.
These prevent rejecting a transform merely because it is not lyric. Polarity may invert: near-death can become restoration, muteness compelled speech, failure hyperfunction; BJ proves inversion cannot by itself disqualify. Stations may redistribute across persons: Addition D and Josephus move fire from the experiencing woman to the sovereign's face and distribute the machine across a dyad. Inventory may shed according to carrier: ears and sweat disappear first in prose while relational sequence survives more deeply; the Suppression Map (#1496) codifies inventory shedding by carrier profile and polarity freedom. Lexical identity is not required: the operator can reduce to a syntactic or relational skeleton; Arabic lexical difference cannot count as negative evidence by itself.
But causal topology is not free. Actors and channels may redistribute; an arbitrary set of bodily words may not be turned into a match. Encounter must actually cause transformation, and the transformation must perform something in the scene. That constraint is where this test either becomes useful or degenerates into motif hunting.
Set K = {5:83, 7:143, 8:2, 39:23, 41:20}: seen before QFO-01; never counted as discoveries. Set K* = each K verse ±3 ayāt, bounded by its sūra: excluded from blind discovery statistics to prevent contextual leakage. Set U = all remaining Qurʾanic text: the actual discovery universe. After QFO-01 is locked, K may be scored exactly like everything else as a contaminated comparison panel. If a K verse looks exciting impressionistically but fails the frozen test, the failure is kept; the operator is not moved toward it. The protocol is specifically designed so that prior excitement cannot determine the result.
Primary Arabic text: the archive's own verbatim seat of Tanzil Uthmani v1.1 (EA-CORPORA-02/08, data/corpora/quran/original/quran-uthmani-tanzil-1.1.txt), 6,236 ayat, license declared in-file, distributed unmodified.
TANZIL_TEXT_SHA256 = bf4f57b968d03f4131c070b1e285da9be0e0a108a21c910e872801ca273312c8 (recorded in the seat's source.json and MANIFEST at commit d2ea2ef4c).
Morphological and root annotation: Quranic Arabic Corpus v0.4 (QAC_DATA_VERSION = 0.4), which builds its annotation over the verified Tanzil text. The run additionally records QFO_PROTOCOL_SHA256 (this deposit's own canonical hash, i.e. its axn_canonical), RUN_TIMESTAMP, and CODE_SHA256 for all retrieval and scoring code.
Translations are navigation aids only. Final scoring is against Arabic plus morphological/syntactic evidence; no candidate can pass because an English translator happened to choose "tremble," "burn," or "see." Tafsīr is quarantined until after candidate scores are locked: commentary must not tell us what a scene "means" before we have measured what its language does.
The nine English semantic buckets M1–M9 are frozen now. An Arabic lexical-mapping stage then receives those buckets but no Qurʾanic concordance and no verse list; it maps each bucket to Arabic roots/lemmas using independent lexical resources and QAC's dictionary, logging every proposed root and every rejection.
The scan has two retrieval arms. Arm A, multi-channel: every ayah hitting at least two distinct material buckets becomes an anchor. Arm B, hard-core: every occurrence belonging to M3 (voice/tongue), M9 (death/collapse), or an unambiguous sensory-occlusion expression becomes an anchor even without a second channel. Each anchor produces exactly one fixed seven-ayah window (anchor ±3), truncated only at sūra boundaries; overlapping windows merge mechanically; any window intersecting K* is removed from U and sent to the contaminated panel.
This answers what the Josephus scan taught: exact co-occurrence is too coarse, and the machine may be narratively distributed (#1493). No semantic-model search is permitted in QFO-01; if lexical retrieval proves inadequate, semantic retrieval can be registered later as QFO-02, so that an embedding model cannot become an invisible second hypothesis-generator.
No immediate collapse into one number. Each candidate receives the vector V = <M9, R5, O, D, P>: nine binary material-station observations; five binary relational-clause observations; topology O; distribution D (single-body / dyadic / multi-agent); polarity P (forward / inverted / mixed).
Topology: O=0 if bodily motifs merely coexist. O=1 if encounter and somatic transformation are linked but later relations are weak. O=2 if a causal sequence runs through reception → alteration/occlusion → threshold or consequential changed state. O=3 if that sequence additionally resolves into renewed or redirected transmission, witness, utterance, reception, or inscription.
The scoring sheet quotes the Arabic locus for every 1. No undocumented intuitive scoring.
Candidates that survive the screen are adjudicated against C1–C12 of #1494, incorporated by reference and unmodified: sight-frame invariance with instantaneity marker (C1); channel-conservation across the dyad with fire on the facing member (C2); voice and death load-bearing, polarity-flippable, absence of both disqualifying (C3); shedding as carrier-diagnostic, ears+sweat first casualties outside lyric (C4); color naming a surface (C5); occlusion as the gate — the seizure purchases something (C6); reactivation-by-another obligatory in narrative, feeding and fire-words marked (C7); agentless grammar at the joints (C8); simultaneity to alternation under instantaneity compression (C9); second-order intensification toward the apparatus where a redactional chain is visible (C10); cycle-formation with the sovereign grant in the tolmaton-slot (C11); thesis-compression formula (C12). Discrepancy rule: where the screen and the C-operator disagree about a candidate, the C-operator governs, and the disagreement is reported, not reconciled invisibly.
Before opening U, a positive-control panel is scored consisting only of independently recognized pre-Qurʾanic receptions of Fragment 31: Catullus 51 certainly; then the published Theocritus 2, Apollonius Argonautica 3, and Valerius Aedituus cases once their exact loci and bibliography are locked from the companion survey (The Transforming Body of Sappho 31, deposited alongside this protocol). A matched negative-control panel of bodily/emotional scenes from the same literary worlds, never proposed as receptions of 31, is scored identically. The positive/negative distributions determine the admissible region before the Qurʾan is run. Known ancient receptions establish what a material reception looks like under this instrument; the question is whether any U candidate falls inside that region. If the instrument cannot distinguish recognized receptions from matched non-receptions, the instrument fails before the Qurʾan gets a chance to "confirm" anything.
After retrieving the U candidates, a matched random sample of noncandidate seven-ayah windows is selected deterministically, using the corpus hash as the random seed, and blind-scored identically. This answers the specificity question: is the configuration unusual in the Qurʾan, or does Qurʾanic discourse produce comparable reception/body/threshold sequences everywhere? Both results are informative. If ubiquitous, the operator as frozen is too coarse to establish a distinctive family in this corpus. If a small number of passages strongly separate from background and resemble the ancient controls in both material inventory and causal topology, that is much stronger.
Candidate windows and controls are shuffled and assigned opaque IDs (QFO-A0001, QFO-A0002, …). Verse numbers and sūra names are hidden during first-pass scoring; the scorer receives Arabic, morphological gloss, and the frozen rubric. Perfect corpus blindness is impossible — the Arabic may be recognizable — but hypothesis blindness is achievable: the scorer is not asked "is this Sappho?", only whether M1–M9 and R1–R5 occur. At least two independent scorers where available; disagreement retained, never reconciled invisibly.
Positive controls fail to separate from negatives → QFO-01 invalid; stop. No U candidates cross the calibrated reception region → no blind Qurʾanic support for P_material under QFO-01. Only K/K* passages score strongly → prior noticed correspondences survive scoring, but no independent discovery. One U passage strongly separates → new material-transform candidate. Multiple independent U passages strongly separate with systematic carrier behavior → evidence of an operative transform-family in the Qurʾanic corpus. The above plus historically independent transmission evidence → begin testing Sapphic reception/lineage, separately, as P_historical.
Multiple Qurʾanic realizations need not be identical: one may drop sweat/hearing, another invert collapse into restoration, another distribute the channels — and that variation becomes evidence if it follows the carrier rules frozen in §4 and C4. It cannot be rescued afterward by inventing a new rule for each miss.
This protocol loses if: the recognized positive controls are not distinguishable from negative controls; U yields no candidates above the frozen control boundary; apparent matches exist only in translations and disappear under Arabic morphological/syntactic inspection; strong scores depend on enlarging windows after seeing results; the supposed sequence dissolves into disconnected motifs; or Qurʾan-internal background windows reproduce the pattern at comparable frequency. Any post-result change to stations, topology, lexical buckets, window size, or pass criteria creates a new protocol version; it does not repair QFO-01.
The deeper virtue of this design is that 39:23 is now allowed to fail. So is 7:143. So is the whole Qurʾan. "I bet this is there" has been converted into a test where the corpus gets to answer. And if it passes — especially if the blind U-set produces passages not yet named — the result is a pre-Qurʾanic operator, calibrated against recognized Sapphic receptions, recovering previously unseen material configurations in an Arabic corpus under frozen carrier rules. That is the experiment.