Practice-based research prototype / 2026 Field open — move, pause, return

The Three
Metamorphoses精神三变

A responsive field where movement, stillness and return become traces—without diagnosing the viewer.

Working Web Fieldauthor-led calibration completed / participant study planned

moveleave a tracepauseform a residuereturnaccumulate

Research dossier

01 / Research context and inquiry

How can changing relations of weight, friction and play become perceptible without defining the viewer?

The project begins from lived experiences of carrying, refusing and beginning again. Nietzsche’s “Three Metamorphoses” gave those experiences a philosophical vocabulary. Practice-based experiments then ask how that vocabulary might become perceptible as a responsive condition without becoming an image, diagnosis or identity label.

Research sequenceContext→Translation→System→Evidence→Limits→Study

Working proposition

Transformation is approached as a changing relation rather than represented as a sequence or assigned as an identity. Weight, Friction and Play are inspectable artistic variables used to construct that relation; they are not claims about a person.

Lived resonance / not autobiography

I had experienced carrying, resisting and beginning again as real changes, but not as clean or dramatic stages. Even rupture appeared gradually, overlapping with responsibility, hesitation, curiosity and the making of new possibilities.

The work does not retell my biography or assume that every visitor shares the same story. Personal experience is the origin of the question, not the content the viewer is asked to recognise.

Philosophical trigger

Nietzsche’s camel, lion and child offered three figures of transformation. This project does not illustrate them literally. It asks what happens when bearing, refusal and beginning are translated into relationships of movement, stillness, repetition and return.

The shift is from “Which stage am I in?” to “How does the field change as I enter it?”

Primary question and supporting inquiries

RQPrimary→

How can changing relations of Weight, Friction and Play become perceptible through interaction without defining the viewer?

SQ1Translation→

How can philosophical qualities be operationalised as behavioural and visual rules without becoming literal symbols?

SQ2Perception→

Which changes in density, interruption, persistence and opening are actually noticed by participants?

SQ3Scale→

What changes when the inquiry moves between screen interaction, a handheld encounter and a proposed bodily space?

Actual development / reconstructed history

The project did not begin as one fully planned technical system. A responsive visual field and an interest in the M5Stack Cardputer developed as adjacent experiments: one concerned traces and changing images; the other concerned holding, entering, waiting and receiving a brief response.

The current research architecture brings them into relation without claiming that they were technically connected from the beginning. This distinction between historical development and present interpretation remains visible throughout the project.

Why a web field

Hand scale / direct trace

The browser prototype isolates the smallest unit of the question: how movement, pause and return produce visible consequence.

Why a threshold

Entry scale / deliberate arrival

The handheld prototype slows the encounter down before the field begins, making entry an intentional act rather than a click.

The field does not tell the viewer what state they are in. It changes with the way they enter it.

02 / Conceptual translation → operational model

Camel, Lion and Child inform three coexisting relations: Weight, Friction and Play.

The philosophical figures are neither illustrated nor assigned to the viewer. They inform a hand-authored analytical model through which behavioural features and perceptual consequences can be compared.

Three coexisting forces / not a fixed sequence

01 / Camel informs

Weight

Burden, endurance, repetition, lingering and accumulation.

02 / Lion informs

Friction

Resistance, contradiction, interruption, displacement and rupture.

03 / Child informs

Play

Exploration, connection, opening, creation and new possibility.

Nietzsche presents a sequence; this artwork deliberately tests a coexistence model because lived change felt overlapping, recurrent and unfinished. This is an artistic reinterpretation, not a universal account of Nietzsche or human transformation.

Rule evidence / input → state change → perceptual consequence

EncounterRuleVisible consequence

MoveMovement deposits an active trace.A path extends and disturbs the existing field.

PauseShort stillness briefly holds the local field.A small residue remains around the last position.

RemainLonger stillness shifts the field from hold to release.The image diffuses, decays and leaves an afterimage.

RepeatSpatial return increases local density and retention.The familiar path becomes thicker and remains longer.

TurnCurvature, abrupt direction and self-intersection increase rupture.Continuity briefly separates, shifts or leaves a gap.

ExploreEntering a low-density or unvisited area increases emergence.Sparse points establish a new local direction.

Operational model / two temporal scales

01 / continuousLocal field responseEach movement segment reads local speed, turn and prior density as accumulation, rupture and emergence

02 / immediateResidue and flowThe local response alters particles, direction, spread, hold and decay while the encounter is happening

03 / completed pathTrajectory analysisA released gesture is resampled for speed, duration, curvature, crossings, repetition, novelty and spatial reach

04 / reflectiveWeight / Friction / PlayThe completed path produces three independent relations that shape echoes, displacement and openings

Technical note / inspect the hand-authored mapping

WWeightincreases with slower movement + prior spatial overlap + longer duration

FFrictionincreases with curvature + abrupt turns + self-intersection

PPlayincreases with novelty + directional continuity + spatial reach + exploratory duration, without automatically saturating

Model statusHand-authored, tuneable and provisional. The three signals describe the current trace under the present rule set. They are not machine-learned predictions, psychological measurements or labels assigned to a person.

Restricted visual primitive library

DustTraceBlurNoiseAfterimageDisplacement

Development trajectory

The project began as a non-interactive particle state. Interaction entered gradually: first through mouse direction and trace, then through optional low-resolution camera sensing as a different kind of movement, rhythm and breath.

Particles, diffusion and water-like movement were chosen because the three forces felt microscopic, elusive and continuously flowing. Brighter halos, decorative light and firework-like effects were tested and rejected: they made transformation appear more violent and theatrical than it felt in lived experience.

01 / Initial stateParticlesNon-interactive matter in continuous transition.

02 / Direction entersMouse traceIntentional movement gives the field a local path.

03 / Body entersCamera presenceCoarse movement and brightness change affect the field as atmosphere.

Visual development / not three metamorphosis stagesParticle state → directional input → embodied input

Practice-based research cycle

01MapTranslate philosophical qualities into behavioural conditions

02BuildDevelop a restricted visual grammar through creative coding

03TuneAdjust parameters through repeated perceptual and embodied tests

04ObserveCompare participant descriptions without giving philosophical labels in advance

05ExpandTranslate the tested interaction from screen to threshold and spatial field

03 / Live field system

Behaviour becomes trace; trace becomes a temporary perceptual field.

The Web Field is the working prototype. It tests how intentional movement, stillness, repetition and optional coarse presence alter density, direction, rupture and decay.

System map / one encounter → two analytical timescales → one field

01 / EncounterMovement, pause and return

Intentional pointer path / optional coarse frame change

02A / During movementLocal relation

Speed / turn / prior density → accumulation / rupture / emergence

02B / After releaseTrajectory relation

Whole-path features → Weight / Friction / Play

03 / Shared consequencePerceptual field

Flow / echo / interruption / opening / afterimage / decay

Primary questionHow can changing relations of Weight, Friction and Play become perceptible through interaction without defining the viewer?

ContributionA philosophical encounter reworked as an open, responsive field—not a symbolic illustration or a model of the viewer.

Working now

Two temporal scales operate in the browser: continuous local field response and completed-trajectory analysis. Both are exported separately so the visual field is not reduced to one score.

Next inquiry

Verify the revised V2 mapping, then observe 6–8 unguided encounters and compare participants’ descriptions before spatial expansion.

Live trace readout / activates after a completed path

Trace tendency / not a personal state

Return to the field, press and draw. The changing relation below describes only how the current rule set reads that path.

No trace measured

Weight

Friction

Play

Two inputs / two kinds of presence

Mouse / intentional movement

Precise, directional, trace-making

Movement establishes direction and fine trace. Repetition builds local density; abrupt changes introduce friction; exploratory paths open unused space.

Optional camera / coarse visual change

Atmospheric, frame-based, less precise

If the visitor actively enables it, low-resolution luminance differences detect coarse movement and brightness change in the camera frame. These signals affect global pressure, breathing and instability differently from the mouse.

Current camera test / reflection

The camera was introduced to test a kind of presence that is atmospheric rather than directional. In the current prototype, larger changes in the frame intensify bright points and can produce an expansive burst.

That response may be too explosive for the project’s quieter logic. The next iteration will test global pressure, breathing and brief instability without treating the camera as another drawing instrument or assigning meaning to the body it senses.

Camera ethicsOpt-in only. A 48 × 36 luminance sample is processed locally and temporarily; no image is displayed, stored or transmitted, and no identity recognition is used. The field remains usable without camera access.

Three optional ways of entering the field

V1 calibration / six controlled actions, one operator

01 / directional openingFast1.35 s · speed 0.785 · Weight 0.252

A brief new path produced little accumulation and a short-lived field.

02 / sustained passageSlow4.78 s · speed 0.129 · Weight 0.630

Longer input built a denser field with stronger persistence.

03 / hold and decayStill12.45 s longest stillness · 37 residues

Local residue appeared briefly, then stabilised, diffused and slowly lost density.

04 / spatial returnRepeat3 paths · return 1.00 · Weight 1.00

The first path opened space; return converted novelty into accumulation.

05 / directional ruptureTurnpeak turn 2.931 · 6 crossings · Friction 0.460

Sharp angular change produced more Friction than a rounded loop, even when both paths crossed themselves.

06 / spatial openingExplore4.73 s · novelty 1.00 · repetition 0.00

A long path through unused space remained open and non-repetitive, exposing V1 Play saturation and prompting the V2 revision.

These P01 hand-run checks calibrate the artefact. Repeat and Turn differentiated clearly; Play saturated across most novel traces, so V2 reduced automatic saturation. These are not participant findings and do not assign meaning, identity or psychological state to a viewer.

04 / Embodied evidence and prototype status

The working field is documented as a relation between body, input, screen and accumulating trace.

Screen capture shows what the field produces; physical documentation shows how it is encountered. Together they make the interaction loop visible without presenting an author-led calibration as a participant study.

Current / embodied record

Body → input → field

The photographs move from the spatial situation to an over-the-shoulder interaction and then to the hand–screen relation. They are evidence that the work operates at a physical scale, not only as an abstract image.

01 / Spatial situation — author, interface and live field in one working environment

02 / Active loop — hand movement and visual feedback remain visible together

03 / Interaction detail — input becomes trace, accumulation and residue

15 s / initial passage

20 s / accumulating trace

25 s / residue and overlap

Evidence boundaryAuthor-led prototype demonstration and calibration record. These images show the interaction situation and temporal change; they are not participant findings or evidence of psychological transformation.

PrototypeRole in the researchStatus

Web Field

Hand-scale test of movement, stillness, repetition and residue.

Working prototype

System calibration

Controlled P01 actions used to inspect the hand-authored mapping and revise V1 Play saturation.

Completed / V2 revised

Physical documentation

Body, hand, interface and temporal field changes recorded during author-led demonstrations.

Completed / current evidence

Optional camera

Coarse presence input; tests atmosphere rather than directional drawing.

Experimental input

Cardputer studies

NOT YET and Blue Oracle test limited input, physical keys, waiting and minimal response.

Documented / standalone

Participant observation

Small first study used to revise rules before technical integration.

Planned / 6–8 people

Spatial integration

Cardputer connection and body-scale installation remain future research directions.

Proposed / not implemented

Earlier prototype archive / Cardputer lineage and scale comparison

Current comparative framework / not the original chronology

01 / Hand

Web Field

The pointer produces precise, directional traces; stillness and return change accumulation.

02 / Handheld

Cardputer studies

Questions, physical keys and waiting test a private temporal encounter. They do not currently control the field.

03 / Body / proposed

Spatial field

Projection, surface and sensing would test whether trace and residue remain legible at whole-body scale.

Current / hand scale

Web Field

The live field, mouse response and optional coarse camera sensing are working experiments. The field remains independently usable without the camera and without the handheld device.

Current / handheld studies

Cardputer interface lineage

The M5Stack Cardputer ADV first entered the project through material curiosity: it was a compact, already assembled computer that could be held, typed into and programmed without building custom hardware. NOT YET and Blue Oracle subsequently tested limited input, physical keys, waiting and minimal response. Interpreting the device as an exhibition threshold is a later development, not its original technical function.

01 / Device and field — two current prototypes in relation

02 / Name entry — a physical gesture of arrival

03 / Variable prompt — one question within a changing sequence

Evidence statusThe photographs and archived program logic document standalone Cardputer studies. Communication between the device and Web Field has not been implemented. The exact screen sequence shown in each photograph still needs to be matched to its corresponding program version.

Documented interaction lineages / do not merge into one invented sequence

01 / NOT YET

Position

INSIDE / OUTSIDE / BETWEEN

02 / NOT YET

Movement

WAITING / MOVING / RETURNING

03 / NOT YET

Destination

KNOWN / UNKNOWN / NONE

04 / NOT YET

Wait + response

The selected movement affects the interval of waiting before one short, pre-written response appears.

Later / Blue Oracle

Name + four questions

A later version begins with a name, proceeds through four questions and returns an open short response after waiting. It remains a standalone study rather than a diagnostic tool.

Conceptual relationship / technically separate today

01Cardputer studiesPrivate handheld interaction / limited input, physical keys, waiting and open response

02WebPerceptual field / movement, stillness, repetition and presence become visual conditions

03Proposed relationA simplified exhibition threshold may later activate the web or spatial field through one local event

Future extension / proposed Cardputer connection and spatial translation

Minimum viable connection / proposed test

01Send one eventThe device sends entry:ready only after the reflective sequence is completed

02Keep it localA local WebSocket carries no name, prompt text or written response

03Change one conditionThe Web Field begins a temporary inhale and lowers its rate of decay

04Expose the boundaryConnection state and latency are documented; failure leaves the field independently usable

Implementation statusProtocol proposal only. It connects the completion of a threshold, not the content of a person’s responses; no interpretation, scoring or response transmission is required.

Future / body scale

Spatial translation

The proposed installation asks whether the current digital and handheld questions can be tested in a physical threshold where whole-body movement replaces pointer movement and residue occupies projection, surface and atmosphere. It is a research proposition, not a completed enlargement of the website.

Translation of scale

01Mouse→ body movement

02Browser screen→ spatial perceptual field

03M5 input→ threshold interface

04Camera presence→ spatial sensing

05Digital residue→ projection / surface / material residue

Three spatial components

01 / Entrance

Threshold Interface

A small, deliberate entry action acknowledges presence before activation.

02 / Transition

Threshold

A membrane, slit or partial opening marks passage without becoming a theatrical portal.

03 / Interior

Perceptual Field

Projection, scrim and exposed structure hold a quiet, incomplete and responsive condition.

Installation overview01 / 04

Speculative Installation Proposal, 2026AI-assisted spatial visualisation developed from the current working prototype.

05 / Limitations, ethics and epistemic boundary

The system describes traces; it does not diagnose people.

The boundary is part of the work. Every sensing and rule-making decision is paired with what the system cannot know, so the field remains open rather than becoming an instrument of classification.

Agency and boundary

The visitor chooses

Whether to move, wait, repeat, draw, enable the camera or leave.

The field receives

Movement, stillness, spatial relation and temporary low-resolution frame differences.

The field changes

Visual density, continuity, instability, diffusion and persistence.

The system cannot know

Identity, intention, emotion, belief or a person’s “true” metamorphosis.

Non-diagnosticNo personal state

Weight, Friction and Play describe relations in the current trace. They do not identify emotion, personality, intention, belief or a viewer’s “real” metamorphosis.

Camera ethicsOpt-in and local

A 48 × 36 luminance sample is processed locally and temporarily; no image is displayed, stored or transmitted, and no identity recognition is used. The field remains complete without camera access.

Operational termsDefinitions stay visible

Repetition means spatial overlap with earlier traces; novelty means entering cells not yet visited in this session. These are working definitions, not psychological facts.

Technical boundarySeparation is stated

The Web Field and M5Stack are currently separate. The proposed connection transmits only an entry-ready event, not names, prompts, answers or interpretations.

Limits / open questions

Screen size, pointer hardware, ambient light and familiarity with drawing interfaces can all change the result. The camera reads coarse luminance differences, so movement and brightness can be confounded, and access is uneven across devices.

Nietzsche’s sequence is culturally and historically situated; this coexistence model is an artistic reinterpretation rather than a universal account of transformation.

Sound is intentionally absent from the current prototype. It might make duration and accumulation more bodily, but it could also overstate a field designed to remain quiet and unresolved. A later comparison should test sparse sonic feedback as a temporal layer rather than assume another medium automatically deepens the work.

The computation is exposed so that it can be questioned. The person remains unresolved.

06 / Calibration completed → exploratory study planned

First inspect the artefact; then observe how people perceive it.

A P01 controlled calibration has been completed to test whether the programmed relations behave consistently. It revised the V1 mapping but did not study an audience. The next stage is a small qualitative pilot asking whether different field behaviours are perceptibly distinct without validating Nietzsche or scoring a person.

Completed calibration / what changed before participant observation

01 / supportedReturn → WeightAcross three repeated paths, return increased from 0 to 1 while Weight rose from 0.314 to 1.000

02 / supportedSharp turn → FrictionAngular turning produced Friction 0.460 versus 0.236 for a rounded looping path

03 / revisedNovelty ≠ automatic Play 1.0Fast, Turn and Explore all saturated in V1, so V2 preserves overlap while restoring numerical variation

04 / clarifiedStillness → hold and decayStillness is documented as a temporal field event rather than assigned an invented trajectory score

Study design / small practice-based pilot

PurposePerceptual differentiationDetermine whether accumulation, rupture, opening and decay are noticed as different field behaviours.

Sample6–8 consenting participantsA small exploratory group used to reveal language, ambiguity and interaction problems; not to support generalisable claims.

ProcedureUnguided + controlled comparisonBegin with a free encounter, then compare selected movement, stillness and return conditions from the same empty field.

MaterialSystem logs + participant languageRecord operational trace features, short screen recordings and post-interaction descriptions without collecting names.

AnalysisCompare, do not diagnoseRelate recurring descriptions and points of confusion to the rule mapping; do not infer personality, emotion or transformation.

Revision criteria

QuestionEvidence soughtPossible revision

DifferenceDo participants notice distinct consequences across pause, turn, repeat and explore?Separate mappings that currently collapse into the same effect.

LegibilityCan participants connect an action with a later residue without being told the rule?Adjust timing, persistence or visual contrast rather than add explanation.

OpennessDo descriptions vary, or does the interface force one predetermined interpretation?Remove labels or effects that over-direct meaning.

RestraintDoes any response become decorative, spectacular or disproportionate to the action?Reduce intensity or remove the unsupported layer.

Run 6–8 unguided encounters

Invite consenting participants to enter the field without philosophical labels in advance. Keep camera use optional and document hardware and context.

Collect short descriptions

Ask participants what changed, what remained and what felt resistant or open. Compare language, differences and ambiguities before relating them to rules.

Verify the V2 mapping

Repeat a short calibration to confirm that Play no longer saturates while Weight and Friction retain their supported behaviour.

Test one local connection

Prototype the M5Stack entry:ready event as a minimal threshold signal. Do not transmit names, written responses or interpretive data.

Translate to space cautiously

Use the revised rules to test projection, threshold and residue at body scale. Treat the installation as a research condition, not a decorative enlargement.

Study statusSystem calibration completed with one operator; participant study not yet conducted. The proposed 6–8 person pilot is exploratory and qualitative. It can guide iteration but cannot establish universal psychological meanings or prove the philosophical model.

The next version should make the rules more legible, not make the viewer more readable.

The ThreeMetamorphoses精神三变

The Three
Metamorphoses精神三变