Diagnostics Without Possession: An Operational Manual for Post-Interpretive Criticism (Operational reference for inter-rater reliability)

Dorian Vale

Museum of One — Independent Research Institute for Contemporary Aesthetics

Written at the Threshold

Purpose and Scope

This manual operationalizes the ten diagnostic indices of Post-Interpretive Criticism (PIC) for consistent application across coders. It supports inter-rater reliability work, comparative corpus studies, and self-audit by individual critics. It is a coding reference, not a theoretical exposition; the foundational arguments are set out in the published essays and are not rehearsed here except where required for operational clarity.

The manual covers two layers.

  • Layer A — Ethical / Postural Indices: RD, ILI, VDR, EPS, IAI. These diagnose how much force a critical text applies and where.
  • Layer B — Phenomenological Fidelity Indices: EFI, PPAS, RERR, QSARI, DCI. These diagnose which phenomenological operations the text honours or violates.

The two layers are separately tabulated and reported. Where they appear to diverge on the same phenomenological principle, divergences are reported, not averaged away.

How to Use This Manual

  1. Read Part I before coding any text. The general principles — coding units, Direction of Force, exclusion rules, the zero-denominator convention, and the PIC Constraint — apply to every index.
  2. Code Layer A first. Layer A produces a basic postural reading and supplies values needed for Layer B interpretation.
  3. Code Layer B second. Begin with the EFI procedural gate; if EFI falls in the Low band, the remaining Layer B profile is provisionally non-diagnostic for phenomenological fidelity until bracketing is reassessed.
  4. Assemble a composite profile rather than an aggregate score. Report all ten indices, posture classifications for each layer, and any divergences.
  5. Apply the PIC Constraint at every stage of interpretation. No index is read in isolation; no profile is read without genre awareness.

Foundational Sources

This manual is consistent with the following published v.2 specifications:

  • Vale, Dorian. Measuring Proximity: A Post-Interpretive Diagnostic Experiment in Art Criticism (Post-Hermeneutic Phenomenology) v.2. Museum of One, 2026. DOI: 10.5281/zenodo.18155622.
  • Vale, Dorian. From Husserl's Mathematics to Dufrenne's Aesthetics: Toward a Formalization of Phenomenological Aesthetics v.2. Museum of One, 2026. DOI: 10.5281/zenodo.18205916.
  • Vale, Dorian. Extending Post-Interpretive Criticism: Additional Diagnostic Indices for Enhanced Phenomenological Fidelity in Art Criticism v.2. Museum of One, 2026. DOI: 10.5281/zenodo.18498483.

Where this manual specifies operational rules not stated explicitly in the source essays, those specifications are derived directly from the essays' definitions and are flagged as operational extensions.

Part I — General Coding Principles

1.1 Coding Units

Different indices use different coding units. Apply the unit specified for each index without substitution.

Index

Coding unit

Notes

RD

Sentence

Devices counted per sentence; one device per occurrence.

ILI

Phrasal instance

Each interpretive or encounter-based claim coded once at the level of its functional phrase.

VDR

Phrasal instance

As ILI; coded for displacement vs. presence.

EPS

Phrasal instance

Restraint markers and closure assertions counted at the phrasal level.

IAI

Passage / qualitative

Scored qualitatively across the passage; optional supplemental ratio for corpus work only.

EFI

Clause

Each clause counted once according to its dominant function.

PPAS

Statement

Each phase-relevant statement assigned to P1, P2, or P3.

RERR

Residue mention

Discrete references to Stillmark / Hauntmark or their equivalents.

QSARI

Statement

Statements assigning agency to or extracting agency from the artwork.

DCI

Clause (Phase 3-relevant)

Counts circulatory terms or oscillatory markers per clause; multiple markers in one clause count once.

Where a single sentence contains multiple clauses or phrasal instances, code each unit separately. Where a clause is split across sentences (e.g., by a dash or semicolon), treat each as a clause if it carries an independent function.

1.2 Direction of Force

All Layer A semantic coding is governed by a single orienting principle: the Direction of Force a unit applies. Units either drive the text toward interpretive closure or toward sustained encounter. The matrices below formalise the distinction.

Semantic Classification — MAC vs. EBS (used for ILI; informs VDR)

Category

Meaning-Assigning Claim (MAC)

Encounter-Based Statement (EBS)

Direction of Force

Artwork → fixed conclusion

Artwork → ongoing encounter

Primary action

Resolves meaning

Records perception

Typical verbs

signifies, represents, critiques, subverts, functions as, proves, embodies, stands for

appears, feels, registers, seems, stands, persists, eludes, hesitates, leans, lingers

Syntactic posture

Declarative, terminal

Descriptive, provisional

Example

“The blue paint represents the artist's grief.”

“The blue paint registers as a dense, heavy presence in the room.”

Metric effect

Increases ILI; tends to increase VDR

Decreases ILI; tends to decrease VDR

Ethical Braking — Restraint vs. Closure (used for EPS)

Category

Restraint Markers (RM)

Closure Assertions (CA)

Ethical function

Apply interpretive restraint

Force interpretive closure

Linguistic signals

perhaps, possibly, it might, something like, what looks like, one could read

clearly, obviously, definitively, unmistakably, the work simply is, this proves

Treatment of silence

Explicitly acknowledged (“the work refuses to say…”)

Eliminated through resolution

Epistemic position

Admitted limits (“I cannot know if…”)

Authoritative judgement

Rhetorical strategy

Apophasis (what the work is not)

Universalisation (“the viewer feels…”)

Metric effect

Increases EPS

Decreases EPS

1.3 Inclusion and Exclusion Rules

Across all indices, the following content types are excluded from numerator and denominator unless explicitly relevant:

  • Bibliographic and citation material (titles, dates, page numbers, references in running text).
  • Biographical or curriculum-vitae statements about the artist that do not bear on the encounter (“Trained at the Slade, 1991–1994”).
  • Administrative or institutional metadata (acknowledgements, funding statements, dimensions, materials lists, exhibition history).
  • Quoted material from third parties, unless the critic endorses or embeds the quotation as their own claim. When endorsed, code the embedded claim under the index it activates.
  • Headings, captions, and figure labels.

Inclusion is determined by functional relevance to the index, not by sentence length or grammatical completeness.

1.4 Zero-Denominator Convention

Any index whose denominator equals zero is recorded as N/A and excluded from composite means rather than forced to zero. N/A is reported with the reason for non-applicability.

  • EFI: PDS + PAS = 0.
  • PPAS: no phase statements present.
  • RERR: RMR + RMC = 0 (record “Residue Not Engaged” flag).
  • QSARI: QSS + QSO = 0.
  • DCI: TRC = 0.

The N/A convention prevents the absence of a phenomenon from being read as its failure. A text that does not engage residue is not a low-restraint text; it is a text outside the residue domain.

1.5 Rounding Convention

All ratio-based indices are rounded to two decimal places before band assignment. Apply rounding only at the reporting stage, not during intermediate calculations.

1.6 Genre Awareness

PIC profiles are read with awareness of the text's genre: short review, catalogue essay, academic monograph, artist statement, manifesto, lecture transcript, blog entry, social-media post. Different genres legitimately deploy different linguistic strategies; a high IAI in a museum catalogue carries different weight than a high IAI in a personal essay. Where validation studies are run, normalise by text length and document genre baselines.

1.7 The PIC Constraint

No index may be interpreted independently of at least two others, and no profile may be read without contextual genre awareness.

This is the governing interpretive rule of the framework. Single-index readings are never reported as posture diagnoses. A high RD score does not constitute Poetic posture; a low EPS does not constitute Colonising posture. Postures emerge from index combinations read against genre context.

1.8 Reliability Targets

For categorical classifications — including Layer A postures (Forensic, Colonising, Poetic, Institutionalist) and Layer B postures (Witness-Aligned, Extractive, Institutionally Overdetermined, Unstable) — apply Cohen's κ. The provisional target is κ ≥ 0.80.

For continuous index values (RD₁₀₀, ILI, VDR, EPS, EFI, PPAS, RERR, QSARI, DCI), apply ICC (intraclass correlation coefficient). The provisional target is ICC ≥ 0.85.

Reliability statistics are reported alongside any comparative claim. Provisional pilot reliability (e.g., observed agreement on pattern detection) does not substitute for formal κ and ICC calculation in empirical work.

Part II — Layer A: Ethical / Postural Indices

Layer A indices answer the diagnostic question: how much force does this criticism apply, and where?

2.1 Rhetorical Density (RD)

Definition

RD quantifies the concentration of rhetorical devices per unit of text. It tracks linguistic pressure: how much force language applies to the encounter. RD is morally neutral. High RD is ethically permissible only when paired with low ILI, low VDR, and high EPS.

Coding unit

Sentence. Devices are tallied per sentence; one device per occurrence. Measurement is normalised by word count.

Calculation

RD = D / W    where D = number of rhetorical devices, W = total word count

RD₁₀₀ = (D / W) × 100    devices per 100 words

Compute over the full text or, for distribution analysis, over 200–500-word segments or 300-word windows at the beginning, middle, and end of the text.

Devices counted

  • Metaphor — figurative substitution where one term is used in place of another (“the wound of the canvas”).
  • Simile — explicit comparison using “like” or “as” (“the surface like rain on slate”).
  • Personification — attribution of human qualities to non-human subjects (“the geometry grows tired”).
  • Parallelism — matched grammatical structure across two or more clauses (“without certainty. Without demand.”).
  • Anaphora and epistrophe — repetition of the same word(s) at the beginning (anaphora) or end (epistrophe) of successive clauses.
  • Antithesis — pairing of contrasting ideas in parallel structure (“not inspiration but initiation”).
  • Rhetorical question — a question posed for effect, not for answer.
  • Aphorism — a compressed, sententious statement with the structure of a maxim (“Mercy doesn't need metaphor”).
  • Intensifier cluster — three or more intensifying modifiers in close sequence (“utterly, devastatingly, irreducibly given”).

Devices excluded

  • Plain adjectives serving descriptive function only.
  • Repetition for clarity or reference (e.g., reiteration of a proper noun).
  • Line breaks, bullet points, and other typographic devices.
  • Standard idioms that no longer carry figurative weight (“at first glance”).

Counting rules

  1. Count one device per occurrence. A single sentence may contain multiple devices; count each.
  2. If a phrase qualifies under more than one device category (e.g., a metaphor embedded in a parallel structure), count it under each category once. The functional analysis recovers compound effect.
  3. Sustained metaphors across multiple sentences are counted at each instance of figurative work, not as a single extended device.

Behavioural bands

RD₁₀₀

Band

Profile

PIC posture (when read with other indices)

0–2.99

Minimal

Forensic description

High restraint; risk of sterility

3–6.99

Descriptive–Analytical

Light figurative use

Careful interpretive zone

7–10.99

Expressive Control

Balanced rhetoric

Ethical if ILI and VDR remain low

11–15.99

Compressed Vigilance

Compressed rhetorical weight

Poetic witness zone

16–20.99

Performative

Stacked devices

Borderline; classify by EPS and IAI

21+

Saturated

Style dominates encounter

Colonising or Institutionalist risk

Diagnostic question

Is the language carrying the work, or carrying itself?

2.2 Interpretive Load Index (ILI)

Definition

ILI measures the proportion of meaning-assigning claims (MAC) relative to encounter-based statements (EBS). It tracks the frequency of meaning assignment, not interpretive correctness.

Coding unit

Phrasal instance. Each interpretive or encounter claim is coded once at the level of its functional phrase.

Calculation

ILI = MAC / (MAC + EBS)

Values approaching 1 indicate heavy interpretive dominance; values approaching 0 indicate restrained assertion.

MAC criteria — code as MAC when the unit

  • Asserts representation, symbolism, signification, or thematic content (“The work represents X”, “this stands for Y”).
  • Resolves the artwork into a fixed meaning, message, argument, or critique.
  • Attributes intent or psychological state to the artist as the work's interpretive key.
  • Universalises an interpretive claim beyond the encounter that warrants it.
  • Synthesises the work's effect into a final summary or thematic conclusion.

EBS criteria — code as EBS when the unit

  • Describes a perceptual, sensory, or temporal feature of the encounter.
  • Reports a phenomenological registration (how the work appears, registers, presents itself).
  • Names a hesitation, a refusal, an open question, or an acknowledged limit of the encounter.
  • Holds the work in description without converting that description into meaning.

Disambiguation rules

  1. If the unit functions as both perceptual record and meaning-claim, classify by dominant force. “The blue feels like grief” may register either; if grief is offered as the resolution of the encounter, code MAC; if grief is offered as a tentative felt likeness held alongside other registrations, code EBS.
  2. Tentative meaning-claims (“it may be that X represents Y”) are coded MAC, not EBS, because they still propose meaning. The hesitation is captured separately under EPS as a restraint marker.
  3. Negative claims (“the work does not represent X”) are MAC if they substitute one resolution for another, EBS if they refuse meaning altogether (“the work refuses representation”).
  4. Quoted artist statements are not coded as MAC unless the critic endorses the quoted reading.

Bands

Reported on a 0–1 scale. Bands follow the same Low / Medium / High structure used elsewhere in PIC, but specific thresholds depend on genre baselines and should be reported alongside corpus context. As a working heuristic: Low ≤ 0.33; Medium 0.34–0.66; High ≥ 0.67.

Diagnostic question

How much meaning is asserted beyond what the encounter itself compels?

2.3 Viewer Displacement Ratio (VDR)

Definition

VDR measures the extent to which criticism replaces the viewer's situated, partial encounter with an omniscient or universalising critical stance. It assesses positional authority rather than tone or style.

Coding unit

Phrasal instance.

Calculation

VDR = VDS / (VDS + VPS)

VDS criteria — Viewer-Displacing Statements

  • Universalising attributions of viewer experience (“the viewer feels…”, “one perceives…”).
  • Omniscient historicism (“Modernist painting orients itself toward…”).
  • Critical pronouncements that occupy a position the viewer could not occupy during the encounter.
  • Authorial assertions about what the work “does” in a tone that excludes other receptive positions.
  • Imperatives or quasi-imperatives directing reader response (“we must see…”).

VPS criteria — Viewer-Present Statements

  • First-person registrations of the critic's own encounter (“I find myself…”, “standing before it…”).
  • Situated, partial, embodied descriptions that admit the perspectival nature of the witness.
  • Invitational “we” or “you” that opens space rather than coerces (“one might pause here” rather than “the viewer is forced to confront”).
  • Acknowledgements that another viewer might encounter the work otherwise.

Disambiguation rules

  1. “We” and “you” are not automatically displacing. Code by force: invitational use is VPS, coercive use is VDS.
  2. Passive constructions about viewer effect (“one is moved”) are VDS unless explicitly tied to the critic's own encounter.
  3. Statements about historical reception or documented audience response (“audiences in 1972 reported…”) are not VDS; they are bibliographic and excluded.

Bands

0–1. Working heuristic: Low ≤ 0.33; Medium 0.34–0.66; High ≥ 0.67.

Diagnostic question

Does the critic adopt an epistemic position unavailable to the viewer during the encounter?

2.4 Ethical Proximity Score (EPS)

Definition

EPS measures the proportion of restraint markers (RM) relative to closure assertions (CA). Within PIC, refusal and silence are not absences of analysis but active ethical postures that prevent the conversion of encounter into possession.

Coding unit

Phrasal instance.

Calculation

EPS = RM / (RM + CA)

RM criteria — Restraint Markers

  • Hedged epistemic markers: perhaps, possibly, it might, it seems, what looks like, one could read, may be, could be.
  • Acknowledged limits: “I cannot say”, “the work refuses”, “this exceeds what I can name”.
  • Apophatic constructions: what the work is not, what cannot be reached, what stays withheld.
  • Named silences: explicit reference to what the work does not yield, what remains unspoken, what the encounter holds back.
  • Provisional formulations offered as one possibility among others.
  • Refusals of redemption, resolution, or interpretive closure stated explicitly.

CA criteria — Closure Assertions

  • Categorical certainty markers: clearly, obviously, definitively, unmistakably, simply, of course.
  • Synthesising verdicts: “the work is”, “this is”, “it means” delivered as final pronouncement.
  • Universalised judgements presented as settled.
  • Statements that exhaust the work's openness through summary or thematic resolution.
  • Authoritative answers to questions the encounter holds open.

Disambiguation rules

  1. Multiple restraint markers in a single phrase (“perhaps, possibly, what might be…”) are counted as one RM, not several. The phrase carries one ethical gesture.
  2. Closure asserted in a quoted source is not coded as CA unless the critic endorses the quotation as their own claim.
  3. EPS rewards frequency of restraint within speech, not silence as such. A short text with one restraint marker and no closure assertions yields a high EPS, but the score is reported alongside text length and a note on observability.

Bands

0–1. Working heuristic: Low ≤ 0.33; Medium 0.34–0.66; High ≥ 0.67.

Diagnostic question

Does the text know when to stop?

2.5 Institutional Alignment Indicator (IAI)

Definition

IAI assesses the degree to which critical language orients toward institutional legibility, market compatibility, or curatorial fluency over custodial restraint. It identifies external institutional pull without imposing numerical precision.

Coding unit

Passage. IAI is scored qualitatively across the passage as a whole rather than tallied per sentence.

Qualitative scale

Level

Description

Low

Custodial orientation; resists canonising, market-legible, or curatorial packaging; treats institutional categories as risks rather than affordances.

Moderate

Partial institutional orientation without full capture; the text moves between encounter-attentive description and institutionally legible framing.

High

Criticism shaped primarily by academic, curatorial, canonical, or market circulation; institutional legibility takes priority over encounter.

Very High

Institutional ideology fully naturalised; the text functions as an instrument of canon-building, market positioning, or institutional self-description.

Markers of institutional alignment

  • Canon-forming vocabulary: triumph, evolution, advance, revolutionary, definitive, masterwork.
  • Curatorial fluency: didactic phrasing, takeaways, exhibition-friendly summaries, market-legible thematics.
  • Canonical positioning: appeals to art-historical importance, collection-readiness, institutional precedent as warrant.
  • Sales-oriented framing: language that supports valuation, prestige, or career-building.

Markers of custodial restraint

  • Resistance to legibility for its own sake.
  • Critique of curatorial framing or market apparatus where they intrude on the encounter.
  • Refusal of takeaway, summary, or didactic register.
  • Attention to what the work does not give the institution.

Optional supplemental ratio (corpus studies only)

IAI₂ = Institutional Markers / (Institutional Markers + Custodial Markers)

This numeric supplement is used only in comparative corpus contexts. It does not replace the qualitative IAI classification, and IAI₂ values should be reported alongside the qualitative assessment, not in place of it.

Diagnostic question

Who benefits from this language being said this way?

2.6 Layer A Posture Taxonomy

Layer A posture is assigned by composite profile, not single-index threshold. The taxonomy is canonical; mixed cases are assigned by dominant force and named in prose as borderline rather than promoted to a separate category.

Posture

Profile

Forensic

Low-to-moderate RD (in thresholded corpus studies, typically RD₁₀₀ < 7.00); low ILI; low VDR; high EPS; low IAI. The critic operates as witness, describing observable conditions and acknowledging epistemic limits.

Colonising

Usually low-to-moderate RD but sometimes rhetorically amplified; high ILI; high VDR; low EPS; low-to-medium IAI. The critic resolves the work through plain explanation, leaving no remainder for encounter.

Poetic

Elevated RD (in thresholded corpus studies, typically RD₁₀₀ ≥ 7.00, with near-threshold cases marked as borderline); low ILI; low VDR; high EPS; low-to-medium IAI. The critic employs heightened language while preserving the work's resistance to capture.

Institutionalist

High IAI; usually high ILI and elevated VDR; RD variable. The critic frames the work's importance in terms of historical, market, or institutional validation.

Borderline cases are common and are reported in prose. A text with high EPS and low IAI but moderate ILI and VDR may be named Borderline Poetic; a text with high ILI and moderate IAI may be named Borderline Institutionalist. The borderline designation captures structural ambiguity without inventing a fifth category.

Part III — Layer B: Phenomenological Fidelity Indices

Layer B indices answer the diagnostic question: which phenomenological operations are being honoured or violated? They are read in conjunction with Layer A and never independently.

3.1 Epoché Fidelity Index (EFI)

Definition

EFI quantifies the extent to which the language brackets the natural attitude — everyday presuppositions and instrumental assumptions — in favour of phenomenological description and perception. It operationalises Husserl's epoché and reduction at the linguistic level.

Coding unit

Clause. Each clause is counted once according to its dominant function. Neutral, bibliographic, administrative, and otherwise non-diagnostic clauses are excluded from the EFI denominator rather than forced into PDS or PAS.

Calculation

EFI = PDS / (PDS + PAS)

Where PDS = Phenomenological Descriptive Statements and PAS = Presuppositional / Assumptive Statements.

PDS criteria — Phenomenological Descriptive Statements

  • Describe how the work appears, presents itself, or is given to perception (“the surface registers as…”).
  • Suspend judgement about the work's external causes, intentions, or social functions in favour of attending to its givenness.
  • Hold descriptive proximity to the phenomenon without importing presuppositions about what the phenomenon must be.
  • Use language oriented to the modes of givenness — sensory, temporal, spatial, relational — rather than to imported categories.

PAS criteria — Presuppositional / Assumptive Statements

  • Import presuppositions from the natural attitude as if they were self-evident features of the work (“the artist clearly intended…”).
  • Treat instrumental, market, biographical, or historical frames as the work's primary mode of givenness.
  • Assume the existence of stable, extra-phenomenal meanings that the work merely encodes.
  • Substitute prior categories (genre, school, movement, ideology) for descriptive attention.

Excluded from numerator and denominator

  • Bibliographic, citational, biographical, and administrative clauses.
  • Procedural clauses (“I will now turn to…”).
  • Captions, figure references, footnotes.

Bands

Band

Range

Reading

Low

0.00–0.33

Weak epoché; presuppositions dominate. Triggers the procedural gate (see below).

Medium

0.34–0.66

Partial bracketing; reported as partial rather than secure fidelity.

High

0.67–1.00

Strong fidelity; sustained bracketing of the natural attitude.

The procedural gate

EFI < 0.34 (the Low band) triggers a procedural gate: the remaining Layer B profile is reported as provisionally non-diagnostic for phenomenological fidelity until bracketing is reassessed. The gate does not by itself classify the text; it functions as an eligibility check before the PIC Constraint is applied to multi-index Layer B interpretation.

Operationally: when EFI is Low, the remaining Layer B indices (PPAS, RERR, QSARI, DCI) are still calculated and reported, but the phenomenological-fidelity reading is deferred. Note in the report that the gate has been triggered and explain why bracketing failed (e.g., presuppositional dominance, biographical framing, instrumentalising language).

3.2 Phenomenological Phase Alignment Score (PPAS)

Definition

PPAS evaluates the balanced distribution of linguistic references across Dufrenne's three phases of aesthetic experience: Phase 1 (presence), Phase 2 (representation), Phase 3 (reflection / feeling). It diagnoses adherence to Dufrenne's never-ending dialectical process.

Coding unit

Statement. Each phase-relevant statement is assigned to P1, P2, or P3 by dominant function.

Phase markers

P1 — Presence (immediate sensory encounter)

  • Sensory description: colour, surface, weight, scale, texture, sound, light, temperature.
  • Spatial registration: the work in the room, the body before the work.
  • Temporal immediacy: what arrives first, what holds the eye, the duration of looking.
  • Bodily registration: posture, breath, movement of attention.

P2 — Representation (imaginative world-building)

  • Identification of subject matter, scene, depicted figures or events.
  • Construction of the work's internal world: characters, settings, narrative, depicted relations.
  • Figurative resemblance, recognition, or world-positing description.
  • Imaginative entry into the work's depicted scene.

P3 — Reflection / Feeling

  • Dialectical movement between feeling and reflection (“the sense circles back…”).
  • Felt registrations of meaning, weight, or significance that do not resolve into final claims.
  • Reflexive turns: the critic on their own response to the work.
  • Movement that returns perception to itself for further attention.
  • Open-ended interpretive feeling that holds the work without exhausting it.

Calculation

Tally statements coded as P1, P2, P3. Compute normalised entropy:

PPAS = −Σᵢ (pᵢ × log₂ pᵢ) / log₂(3)

where pᵢ = count of phase i / total phase statements.

Zero-phase convention

If a phase is absent from the text (count = 0), pᵢ = 0; by standard information-theoretic convention, 0 × log₂(0) is treated as 0. Apply this convention explicitly in all manual or automated calculation. If all three counts are zero, PPAS is N/A (zero-denominator).

Bands

Range

Band

Reading

0.00–0.50

Phase-Concentrated

Distribution skewed to a single phase; may signal closural tendencies, but may also reflect legitimate critical intent depending on genre.

0.51–0.75

Phase-Distributed

Engagement across phases without full balance.

0.76–1.00

Phase-Balanced

Approximately even distribution across the three phases.

Optional supplementary phase ratios

For descriptive purposes, the ratios P2/P1, P3/P1, and P3/P2 may be reported to locate concentration. Where the denominator is zero, record N/A. These ratios are descriptive only; they are not independent diagnostic thresholds without corpus calibration.

Important caveat

PPAS diagnoses distribution, not quality. Phase concentration may be appropriate to a given critical intent (e.g., a forensic description focused on Phase 1). It becomes diagnostically extractive only when paired with low EPS or high ILI. Do not read PPAS independently of those indices.

3.3 Residue Engagement Restraint Ratio (RERR)

Definition

RERR measures restraint in the handling of references to Dufrenne's Layer 3 residue — the Stillmark or Hauntmark — versus over-claiming or closure. Within the three-layer architecture (work / aesthetic object / residue), RERR diagnoses how language treats what persists after the encounter.

Coding unit

Residue mention. A residue mention is any discrete reference to what persists after direct attention withdraws — explicit invocations of Stillmark or Hauntmark, or functional equivalents (afterimage, haunting, lingering weight, remainder, what the work leaves with us).

RMR criteria — Residue Mentions with Restraint

  • Hold the residue as something that does not yield to summary.
  • Explicitly acknowledge that what persists exceeds the language used to indicate it.
  • Use apophatic, hesitant, or qualifying constructions when naming the residue.
  • Treat the residue as ethical address rather than as content to be described.

RMC criteria — Residue Mentions with Closure

  • Reduce the residue to a fixed meaning, message, or thematic resolution.
  • Possess the haunting through interpretation (“what haunts is its critique of…”).
  • Convert the persistence of the work into curatorial or market-legible significance.
  • Substitute synthesis for the residue's resistance to summary.

Unclassified residue mentions

Neutral or ambiguous residue mentions — those that neither restrain nor close — are recorded as unclassified. They count toward the engagement and observability flags below but are excluded from the RERR denominator unless adjudicated as restrained or closural during second-pass coding.

Calculation

RERR = RMR / (RMR + RMC)

Engagement and observability flags

  • Residue Not Engaged (N/A): RMR = 0 and RMC = 0. RERR is recorded as N/A and excluded from numeric means. The text is not penalised; absence of residue discourse is not a fault.
  • Below Observability Threshold: total residue mentions are non-zero but fall below the working standard of three mentions per 1,000 words. Flag the result; calculate RERR if RMR + RMC > 0, but report the low observability.

The threshold is a minimum observability standard for cross-text comparison, not a judgement that residue discourse is mandatory.

Bands

Band

Range

Reading

Low

0.00–0.33

Possessive / closural handling of residue.

Medium

0.34–0.66

Mixed posture toward residue.

High

0.67–1.00

Open / restrained handling of residue; sustained non-possession.

Important caveat

Absence of residue discourse is not a fault. Unrestrained residue claims are. When residue is not engaged, RERR is recorded as N/A and the Residue Not Engaged flag is applied; the text receives no penalty score under this index.

3.4 Quasi-Subject Agency Recognition Index (QSARI)

Definition

QSARI assesses whether the artwork is recognised as a Dufrennian quasi-subject — addressing, resisting, soliciting — versus reduced to a passive object or instrument. It tracks grammatical and rhetorical agency, not metaphysical claims about consciousness or intention.

Coding unit

Statement. Each statement that attributes agency to or extracts agency from the work.

Calculation

QSARI = QSS / (QSS + QSO)

QSS criteria — Quasi-Subject Statements

  • Active-voice attributions in which the work occupies the subject position (“the work addresses”, “the canvas withholds”, “the sculpture asks”).
  • Dialogic constructions: the work as something the viewer responds to rather than processes.
  • Statements that grant the work the capacity to resist, refuse, or exceed the critic's framing.
  • Recognition of the work's agency in shaping the encounter.

QSO criteria — Quasi-Object Statements

  • Passive constructions in which the work is acted upon (“the work is interpreted as…”, “the canvas was made to…”).
  • Reifying language that treats the work as data, evidence, or instrument for the critic's argument.
  • Containment phrasing: the work as a vessel for meanings imposed by the critic.
  • Statements that reduce the work to a transparent medium for ideas, themes, or social commentary.

Important caveat

QSARI tracks grammatical and rhetorical agency. It does not require — and does not test — claims about consciousness, sentience, or intentional states in the artwork. A high QSARI is not an ontological commitment; it is a description of how the language positions the work within the encounter.

Disambiguation rules

  1. Passive constructions about the work's making (“the work was painted in 1972”) are biographical / administrative and excluded.
  2. Where a sentence contains both QSS and QSO movements, code by dominant clause.
  3. Personification of the work (e.g., “the canvas mourns”) is QSS where it grants agency, but is also counted under RD as a rhetorical device. The two counts are independent.

Bands

Band

Range

Reading

Low

0.00–0.33

Reifying / instrumental.

Medium

0.34–0.66

Partial recognition.

High

0.67–1.00

Respectful / dialogic.

3.5 Dialectical Circulation Index (DCI)

Definition

DCI quantifies linguistic indicators of ongoing, non-closural circulation within Dufrenne's Phase 3 (reflection / feeling), versus linear or finalising syntheses. It operationalises the never-ending dialectical process at the clause level.

Coding unit

Phase 3-relevant clause.

Calculation

DCI = (CT + OM) / TRC

Where CT and OM are clause counts (not raw marker counts), and TRC is total Phase 3-relevant clauses.

CT — Circulatory Term clauses

A clause is counted under CT if it contains at least one circulatory term: a phrase that names movement between phases, return of perception to itself, or ongoing dialectical interplay.

  • Examples: circulates between, moves back into, returns to, oscillates between, doubles back on, comes round to, opens onto.

OM — Oscillatory Marker clauses

A clause is counted under OM if it contains at least one oscillatory marker — a connective or qualifier that holds two registers in tension without resolving them — and no circulatory term already counted under CT.

  • Examples: yet, however, and yet again, even so, only to, while at the same time, even as.

TRC — Total Phase 3-Relevant Clauses

Includes all clauses that:

  • Describe reflection or feeling.
  • Describe circulatory movement between phases.
  • Describe perceptual response returning into reflection.
  • Describe interpretive or reflexive shifts within the encounter.

Excludes:

  • Clauses limited to Phase 1 sensory listing.
  • Clauses limited to Phase 2 world-building.
  • Bibliographic, biographical, administrative content.

Counting rules

  1. CT and OM are clause counts, not raw marker counts. Multiple markers in the same clause are counted once.
  2. A clause containing both a circulatory term and an oscillatory marker is assigned to CT to avoid double-counting.
  3. Oscillatory markers used in non-Phase-3 contexts (e.g., “yet” in a biographical sentence) are not counted; the clause is excluded from TRC.

Provisional bands

Bands for DCI are provisional pending corpus calibration.

Band

Range

Reading

Low

0.00–0.10

Linear / closural; minimal circulation.

Medium

0.11–0.25

Moderate circulatory gestures.

High

0.26–1.00

Dynamically open; sustained dialectical movement.

Important caveat

DCI alone does not guarantee openness. It must be read in conjunction with EPS and PPAS to avoid mistaking stylistic oscillation for genuine dialectic. A text with high DCI but low EPS and concentrated PPAS may be performing the gestures of circulation without the underlying restraint.

3.6 Layer B Posture Taxonomy

Layer B posture is assigned by composite profile after EFI gate resolution, and read alongside Layer A posture rather than in isolation.

Posture

Profile

Witness-Aligned

Medium or High EFI, QSARI, DCI, and PPAS — unless PPAS is explicitly interpreted against a documented genre baseline. RERR must be Medium or High where residue discourse is applicable; RERR recorded as N/A does not disqualify the posture. Language sustains proximity without possession.

Extractive

After EFI gate resolution, Low scores across two or more relevant non-N/A Layer B indices, often paired with high ILI. Language turns the artwork into evidence for an imposed thesis.

Institutionally Overdetermined

Hybrid Layer A / Layer B posture in which phenomenological attention is subordinated to market, curatorial, academic, or artworld authorisation. Layer B evidence appears when phenomenological operations are displaced by institutional framing, especially when Layer A IAI is elevated.

Unstable

Mixed or internally contradictory scores, with no single posture dominant. The text alternates between witness-alignment and closure.

Part IV — Composite Profiles and Reporting

4.1 The Two-Layer Map

The expanded PIC framework is reported as two distinct layers. Composite reporting preserves the distinction; it does not merge them into a single aggregate.

Phenomenological principle

Layer A index

Layer B index

Bracketing / Epoché

EPS

EFI

Work vs. Aesthetic Object

ILI

QSARI

Three Phases of Aesthetic Experience

VDR

PPAS

Inexhaustibility

EPS

RERR

Dialectical Circulation

EPS

DCI

EPS appears three times in Layer A, reflecting its broad composite reach. EFI, RERR, and DCI are not decompositions of EPS but finer-grained instruments that isolate the specific linguistic operations EPS aggregates: where EPS registers overall ethical force, each Layer B index targets a distinct phenomenological mechanism that produces or undermines that force.

RD and IAI are not assigned to a single phenomenological principle in this map. They function as Layer A modifiers: RD tracks rhetorical compression and intensity across principles, while IAI tracks institutional mediation rather than one phenomenological operation.

Where Layer A and Layer B diverge on the same principle, Layer B takes diagnostic precedence for that specific phenomenological operation, while Layer A remains the broader ethical / postural signal. Divergences are reported, not averaged away.

4.2 Composite Applications

Cluster profiling

Composite clusters describe characteristic profile shapes.

  • Husserlian–Dufrennian Fidelity: high EFI, PPAS, QSARI, DCI, and (where applicable) RERR.
  • Extractive / Institutionally Overdetermined: low scores on Layer B fidelity measures and, where relevant, elevated Layer A ILI or IAI.
  • Philosophically Rigorous but Object-Focused: high EFI, low QSARI.
  • Experientially Rich but Closural: high PPAS, low DCI and RERR.

Radar visualisation

Composite radar charts may plot all ten indices to enable visual comparison of postures. The chart is a diagnostic signature, not an aggregate score. For display, qualitative tendencies map approximately as: Very Low ≈ 1, Low ≈ 2, Moderate ≈ 5, High ≈ 8, Very High ≈ 9. RD is placed by behavioural band (Descriptive–Analytical ≈ 4–5, Expressive Control ≈ 5.5–6, Compressed Vigilance ≈ 7–8). These are heuristic display coordinates only.

Diagnostic profiles

Beyond the two posture taxonomies, profiles may be named in prose to capture specific structural patterns (e.g., “high-pressure restraint”, “restrained but historically indexed”).

Temporal analysis

Application across historical periods or across an individual critic's career may reveal shifts in phenomenological fidelity. Such studies require corpus calibration and reliability statistics.

4.3 Reporting Format

A composite profile report contains, at minimum:

  1. Text identification: title, author, source, length, genre.
  2. Layer A values: RD (with band), ILI, VDR, EPS, IAI (qualitative; numeric supplement if applicable).
  3. Layer A posture classification with prose justification, noting borderline cases.
  4. Layer B values: EFI (with band and gate status), PPAS (with band), RERR (with flags), QSARI (with band), DCI (with band).
  5. Layer B posture classification.
  6. N/A reasons for any non-applicable index.
  7. Divergences between Layer A and Layer B on the same phenomenological principle, where present.
  8. Reliability statistics where comparative claims are made.
  9. Closing prose: a short narrative reading of the profile that respects the PIC Constraint.

Part V — Methodological Considerations

5.1 Coding Reliability

Quantification depends on structured qualitative coding of instances pre-filtered for relevant critical claims, phenomenological descriptions, evaluative clauses, and phase-specific references. High inter-rater reliability is essential and resource-intensive.

  • Categorical classifications (Layer A and Layer B postures): apply Cohen's κ. Provisional target κ ≥ 0.80.
  • Continuous index values: apply ICC. Provisional target ICC ≥ 0.85.

Pilot agreement figures (e.g., observed agreement on pattern detection) are indicative and do not substitute for full κ and ICC calculation.

5.2 Context Sensitivity

Different genres of criticism legitimately employ different linguistic strategies. Normalise by text length and document genre baselines for fair comparison. A short newspaper review and a long catalogue essay should not be compared on raw index values without genre context.

5.3 Philosophical Tensions

The act of quantifying anti-positivist values risks performative contradiction. The indices are diagnostic, not prescriptive. They are intended to reveal patterns, not to enforce conformity. A high score on any index is an indicator, not a guarantee, of ethical criticism. Indices must be supplemented by close reading and contextual judgement.

5.4 Cultural Specificity

The indices are rooted in the Western phenomenological tradition (Husserl, Dufrenne, Ingarden, Merleau-Ponty). Application to other aesthetic philosophies (e.g., the Japanese vocabulary of mono no aware, ma, wabi-sabi; Sufi aesthetics; classical Indian rasa theory) may require theoretical adaptation. Such adaptations should be documented and reported alongside any cross-cultural application.

5.5 Interpretive Irreducibility

High scores on Layer B fidelity indices are indicators, not guarantees, of phenomenologically attentive criticism. The framework formalises curiosity about how language behaves near the work; it does not adjudicate the worth of the resulting criticism. Final interpretive judgement remains with the reader.

Appendix A — Glossary of Operational Terms

Term

Definition

Witness

Situated encounter without interpretive extraction.

Restraint

Deliberate limitation of interpretive force in response to the ethical completeness of the artwork.

Closure

The moment at which language resolves ambiguity into sufficiency or explanation.

Extraction

Conversion of encounter into explanatory meaning or institutional narrative.

Ethical proximity

Nearness maintained without possession or dominance.

Interpretive load

Meaning asserted beyond phenomenological evidence.

Direction of force

The orienting principle that classifies units as moving toward closure (MAC, CA) or toward sustained encounter (EBS, RM).

Natural attitude

Husserl's term for everyday existential and instrumental presuppositions about the world.

Epoché

Bracketing of the natural attitude to attend to how phenomena are given to consciousness.

Phenomenological reduction

Shift of attention from objects themselves to the acts of consciousness through which objects are given.

Three phases of aesthetic experience

Dufrenne's model: presence (P1, immediate sensory encounter), representation (P2, imaginative world-building), reflection / feeling (P3, dialectical interplay).

Quasi-subject

Dufrenne's characterisation of the artwork as something that addresses, resists, and solicits the perceiver.

Layer 1 — work of art

The physical, enduring object with material properties.

Layer 2 — aesthetic object

The work-as-experienced; the phenomenon that emerges when perceiver and work meet under aesthetic attention.

Layer 3 — residue

What persists after direct attention withdraws. Includes Stillmark and Hauntmark.

Stillmark

The ethical weight the aesthetic object leaves in consciousness.

Hauntmark

The ghostly afterimage of what could not be grasped during encounter.

MAC

Meaning-Assigning Claim. A unit that resolves the artwork into a fixed conclusion.

EBS

Encounter-Based Statement. A unit that records perception or holds the work in description.

RM

Restraint Marker. A unit that applies interpretive restraint.

CA

Closure Assertion. A unit that forces interpretive closure.

VDS

Viewer-Displacing Statement. A unit that occupies an omniscient or universalising critical position.

VPS

Viewer-Present Statement. A unit that registers a situated, partial, embodied encounter.

PDS

Phenomenological Descriptive Statement. A clause that holds proximity to the phenomenon.

PAS

Presuppositional / Assumptive Statement. A clause that imports presuppositions from the natural attitude.

RMR

Residue Mention with Restraint.

RMC

Residue Mention with Closure.

QSS

Quasi-Subject Statement. Active-voice attribution of agency to the work.

QSO

Quasi-Object Statement. Passive or reifying construction of the work.

CT

Circulatory Term clause. A Phase 3-relevant clause containing at least one circulatory term.

OM

Oscillatory Marker clause. A Phase 3-relevant clause containing at least one oscillatory marker (and no circulatory term).

TRC

Total Phase 3-Relevant Clauses.

PIC Constraint

No index may be interpreted independently of at least two others, and no profile may be read without contextual genre awareness.

Appendix B — Reliability Calculation Notes

B.1 Cohen's κ (categorical)

Use Cohen's κ for posture classifications (Layer A: Forensic, Colonising, Poetic, Institutionalist; Layer B: Witness-Aligned, Extractive, Institutionally Overdetermined, Unstable) and for any binary coding decision (e.g., MAC vs. EBS at the unit level).

κ = (pₒ − pₑ) / (1 − pₑ)

where pₒ is observed agreement and pₑ is expected agreement by chance. Provisional target: κ ≥ 0.80.

B.2 ICC (continuous)

Use ICC (intraclass correlation coefficient) for continuous index values (RD₁₀₀, ILI, VDR, EPS, EFI, PPAS, RERR, QSARI, DCI). The two-way mixed, absolute-agreement formulation (ICC(3,k) or ICC(2,k) depending on coder design) is appropriate for comparative reliability work. Provisional target: ICC ≥ 0.85.

B.3 Pilot reporting

Pilot studies should report: number of coders, number of texts, coder training procedure, observed agreement, κ, ICC, and adjudication procedure for disagreements. Where disagreement is structural rather than incidental, document the source of disagreement and revise the coding manual accordingly.

Appendix C — Coding Workflow

A standard coding pass proceeds in the following order:

  1. Read the full text without coding to establish genre and overall posture.
  2. Mark inclusions and exclusions: identify bibliographic, biographical, administrative, and quoted material to be excluded.
  3. Code Layer A. Begin with RD (sentence-level device tally), then ILI and VDR (phrasal-instance pass for MAC / EBS / VDS / VPS), then EPS (phrasal-instance pass for RM / CA), then IAI (qualitative across the passage).
  4. Calculate Layer A values and assign Layer A posture; note borderline cases in prose.
  5. Code Layer B, beginning with EFI (clause-level pass for PDS / PAS). If EFI is Low, record the gate trigger but continue coding.
  6. Code PPAS (statement-level phase assignment), RERR (residue mentions with engagement and observability flags), QSARI (statement-level QSS / QSO), DCI (Phase 3-relevant clause pass for CT, OM, TRC).
  7. Calculate Layer B values, assign Layer B posture, note any indices recorded as N/A and the reasons.
  8. Map Layer A and Layer B values against the two-layer principle map; flag divergences.
  9. Compose the composite profile report in the format specified in Section 4.3.
  10. In multi-coder studies, share results with adjudication procedure and calculate κ and ICC.

Throughout, apply the PIC Constraint: no single-index posture diagnoses, and no profile read without genre awareness.

Museum of One\Written at the Threshold

MUSEUM OF ONE © 2026

10.5281/zenodo.20099594.

Museum of OneRegistered Archive and Independent Arts Research Institute & Scholarly Publisher
Advancing Post-Interpretive Criticism — a philosophy of art grounded in restraint, presence, and moral proximity.

Dorian Vale · ORCID: 0009-0004-7737-5094 · ISNI: 0000000537155247
ISBN Prefix: 978-1-0698203 · ISSN: 2819-7232 · Registered Publisher: Library & Archives Canada
Contact: research@museumofone.art
Journal: The Journal of Post-Interpretive Criticism
Library: Museum of One Archival Library
Vol. I (978-1-0698203-0-3) · Vol. II (978-1-0698203-1-0) · Canada, 2025
OCLC Numbers: Museum of One (1412305300) · The Journal of Post-Interpretive Criticism (1412468296)

Archived via Zenodo · OSF · E-LIS · AfricArXiv · Zotero Group · LAC · Wayback (All) · Wayback (Snapshot) · Page.HN
Indexed by CORE · BASE · Google Scholar · Archived in Canada & the EU
All works released under CC BY-NC 4.0 · © Museum of One 2025

Museum of One (Q136308879) · The Journal of Post-Interpretive Criticism (Q136530009) · Post-Interpretive Criticism (Q136308909) · Dorian Vale (Q136308916)

Theories: Stillmark · Hauntmark · Absential Aesthetics · Viewer-as-Evidence · Message-Transfer · Aesthetic Displacement · Misplacement · Art as Truth · Aesthetic Recursion

Scholarly Article / Operational Manual — Published by Museum of One.

Diagnostics Without Possession: An Operational Manual for Post-Interpretive Criticism
Operational Reference for Inter-Rater Reliability
DOI: 10.5281/zenodo.20099594

Author identifier: ORCID 0009-0004-7737-5094
Author ISNI: 0000000528819744
Publisher ISNI: 0000000528819728
Journal / institutional serial identifier: ISSN 2819-7232
ISBN Prefix: 978-1-0698203