Shoggoth.net The Decoherent Shoggoth.net

A multi-observer entity for Cthulhu Eternal

Incident 6-14: Bellweather Pumping Station

All four witnesses agreed on this much. At 21:17 something came into the eastern service corridor. It stopped under the emergency light. It covered twelve metres in about four seconds. It struck maintenance supervisor Imran Patel once, hard enough to put him through the wired-glass door, and then went on into the turbine hall, where it stayed until electrical engineer Lena Cho brought up the inspection array. Everything past that depended entirely on who you asked.

Lena Cho. It was an animal. Black, six-legged, no hair. The front legs were shorter, so its shoulders kept falling forward. Long skull, two mouths, one over the other. When it hit Imran it bit him round the chest and threw him sideways.

Imran Patel. There was no animal. It was a burned man crawling backward with his stomach to the ceiling. He grabbed my coveralls. His hands were too hot to touch. I know what Cho says she saw. She is wrong.

Officer Daniel Reeve. It looked like a break in the building. Not a crack on anything, a crack standing upright where there wasn’t a surface. The corridor folded along it. Patel was already flying before it reached him.

Paramedic Sofia Alvarez. A sheet of insects came round the corner. Not a swarm, a flat sheet, one insect thick. When it touched the door the glass bent without breaking, and then it broke, after Patel went through it.

Body-camera review. The recording shows a loose bundle of wet electrical cable moving with nothing holding it up. No mouth, no hands, no fracture, no insects. At 21:17:11 the bundle changes position between frames without crossing the distance between. At 21:17:13 Patel accelerates sideways into the door.

Every witness identified the recording as authentic. Every witness insisted it showed the wrong creature.

One organism, several facts

A shapeshifter turns one form into another, and if you are lucky you film the change. The Decoherent never changes at all inside any single observer’s experience. Cho saw the same six-legged animal before, during, and after. Patel saw the same burned man the whole time. Reeve saw the same standing fracture. The body camera recorded the same cable mechanism in every frame.

It is not hallucination either, because each version does real physical work on its own observer. Cho can photograph bite marks. Patel’s coveralls carry hand-shaped burns. Reeve’s laser survey shows a genuine discontinuity in the corridor’s dimensions. Alvarez recovers insect-like fragments. Every one of those evidence trails is locally consistent and repeatable. The contradiction only surfaces when you lay records from different observers side by side.

Underneath the projections there is one organism. It sits in one effective place, performs one underlying action, and has one finite reserve of energy. The forms cannot agree on anatomy, scale, locomotion, or method. They agree completely on the conserved consequences: when contact happened, how much momentum changed hands, how much energy went into the room, and whether the thing was hurt. Shoot the animal in its lower jaw and Patel watches a hole open through the burned man’s sternum, Reeve sees one branch of the fracture die, the camera catches a cable snapping, and the Decoherent loses the same Hit Points in every account. Nobody ever sees the real wound, because there is no observer-independent wound to see in curved space.

What decoherence actually does

Quantum decoherence gets described sloppily as “observation making reality choose,” and that phrasing does two kinds of damage at once. It hands consciousness a magic power it does not have, and it implies that ordinary people looking at the same object should routinely spin up incompatible realities. Neither thing is in the physics.

A quantum system $S$ gets entangled with an apparatus or environment $E$ through plain physical interaction. In a stripped-down measurement model,

$\left(\sum_k c_k\lvert s_k\rangle\right)\lvert E_0\rangle \longrightarrow \sum_k c_k\lvert s_k\rangle\lvert E_k\rangle$

Once the environmental states $\lvert E_k\rangle$ are effectively distinguishable, a local observer can no longer get at the interference between alternatives, and the system’s reduced state goes approximately diagonal in one stable basis, the pointer basis, picked out by the interaction. The environment also spreads the news around. Scattered photons, sound, stirred air, instrument readings, all of it carries redundant copies of the same macroscopic fact, so different observers can sample different fragments and still agree the table is in one place. Quantum Darwinism is the study of how that redundant encoding is what makes classical facts practically objective. So decoherence, done honestly, explains why people agree. It does not hand out private monsters. For that you need an explicit fictional departure, stated plainly.

The Tindalosian axiom

An angular organism has no single stable observable that every curved-space interaction agrees to select. Its coupling to each observer $O_i$ picks out an observer-relative basis $\{\lvert s_k^{(i)}\rangle\}$ :

$H_{\mathrm{int}}^{(i)} = \sum_k \lvert s_k^{(i)}\rangle\langle s_k^{(i)}\rvert \otimes B_k^{(i)}$

For two observers, the observables those couplings select need not commute:

$[A_i,A_j]\ne 0$

So there is no shared basis in which both observers’ projected descriptions can be definite at once. Each local interaction decoheres in the ordinary way. What fails to happen is the usual final step, where one redundantly recorded pointer state emerges across the whole environment. None of this needs a macroscopic superposition and no conscious mind is choosing the appearance. Eyes, cameras, animals, microphones, pressure plates, thermal sensors all count, because each one physically correlates with the thing, and each coupling gives a stable local answer to a different physical question. What you have is one angular process being forced into several mutually incompatible classical descriptions, and from its side of it, every additional observer is one more anatomy nailed onto the same event.

What stays invariant

Set the invariants before play. Without them, contradictory descriptions turn into permission to improvise anything and the monster stops having rules. These properties are shared across every projection: the event location (observers agree on the region, within ordinary measurement error), the event order (one underlying action came before another, unless the scenario separately drags in angular causality), the momentum transfer (the victim is thrown one way with one approximate force, whatever the observers think hit them), the energy deposition (heat, fracture, sound, and electrical disturbance balance globally even when the local forms disagree), the damage (one HP total, and an injury in one projection lands in all of them), the will (one WP total, one decision per action), and the purpose (every projection is the same suffering organism trying to cut down the incompatible observation and get loose).

What is allowed to differ: anatomy and limb count, texture and colour and temperature and smell, apparent size as long as the effective target area holds steady, method of locomotion, the description of attacks and wounds, which mundane trick appears to repel it, and whether it reads as biological, mechanical, human, or geometric. That split is the whole engine of the thing. Describe the contradictions freely; resolve the consequences once.

Observation Profiles

When a Protagonist first directly observes the Decoherent, secretly assign or roll one Observation Profile. That observer keeps it for the whole incident unless their coupling is deliberately replaced by a shared measurement system. The profiles all describe the same statistics; their apparent countermeasures are basis suppressors that protect that observer or weaken that projection without doing a thing to the underlying organism.

1. The Black Animal

A six-legged, hairless animal with a falling gait and a jaw split top to bottom. It smells of hot dust and its body drinks light without throwing a highlight. Invariant Contact appears as: a bite and a violent sideways wrench. Apparent evidence: punctures, saliva-like blue residue, six asymmetrical tracks. Basis suppressor: saturating white light. While the observer keeps it brightly lit, attacks against that observer suffer −20%.

2. The Burned Reversal

A human figure crawling on its back with every joint bent the wrong way. The skin is heat-damaged, but the air around it goes cold. Invariant Contact appears as: a two-handed grab that leaves a thermal injury. Apparent evidence: hand-shaped burns, fibres crushed as if gripped, human skin under the victim’s nails. Basis suppressor: continuous speech aimed straight at it. While the observer talks without a break, they gain +20% to resist Observer Isolation.

3. The Standing Fault

A branching discontinuity with no surface holding it up. Background objects are displaced across it, and it moves by closing at one end and opening at the other. Invariant Contact appears as: space folding across the victim and throwing them along the fold. Apparent evidence: survey errors, laser discontinuities, planar fractures, impossible parallax. Basis suppressor: a stable geometric reference grid. Attacks made through a calibrated sight or surveyed coordinate system ignore 3 points of Consensus Armour.

4. The Arthropod Sheet

Thousands of flat, interlocked arthropods moving as one flexible surface. Nothing ever leaves the sheet, and their legs make no sound until the observer closes their eyes. Invariant Contact appears as: the sheet wrapping a limb or torso and pulling tight. Apparent evidence: segmented fragments, chitin-like dust, parallel abrasions, insect proteins that never show up in later tests. Basis suppressor: moving air. Industrial fans, compressed gas, or a sustained wind impose −20% on attacks against this observer.

5. The Cable Saint

Cables, tendons, and translucent tubes hung around an empty human outline, with electrical connectors where a face should be. Nearby machinery behaves as if it is receiving valid control signals. Invariant Contact appears as: a cable driving into equipment or clothing and throwing an actuator hard. Apparent evidence: electrical faults, punctured insulation, valid but undocumented control packets, conductive blue gel. Basis suppressor: electrical isolation. While the observer and their gear are cut off from conductive structures and live networks, attacks against them suffer −20%.

6. The Occlusion

The observer does not see an object. They see a region with the meaningful detail missing. Eyes skate across it, cameras focus on the background, and people standing behind it get hard to remember. Invariant Contact appears as: the victim vanishing for a fraction of a second and coming back already moving. Apparent evidence: dropped frames, missing words in transcripts, blank patches in drawings, wounds with no remembered cause. Basis suppressor: continuous touch on a fixed reference object. While holding that contact, the observer cannot be surprised by the entity and gains +20% to Alertness tests involving it.

Duplicate profiles

Two observers can draw the same profile, and their agreement is real: similar couplings picked similar pointer states, so they can corroborate each other and coordinate normally. That raises Consensus by 1 if they make a durable shared record. Agreement is useful, and it also forces the thing toward a more definite curved-space body.

The Consensus Track

Consensus measures how much observer-independent structure has been forced onto the Decoherent. Five incompatible memories create less of it than two calibrated instruments measuring the same quantity over and over.

Rating	Common record	Effect on the Decoherent
0: Private states	Observers agree only that something happened.	Each projection is isolated. Hard to target; disengages easily.
1: Shared event	Time and rough location agree.	Attacks can be coordinated against its region.
2: Shared invariants	Momentum, energy, movement, or timing is measured consistently.	Damage through one projection shows up instantly in all the others.
3: Common outline	A repeated procedure yields one stable silhouette or target boundary.	Consensus Armour falls to 3. Basis Slip gets expensive.
4: Common dynamics	Observers predict its next action from a shared model.	It cannot surprise prepared observers; all attacks on it gain +10%.
5: Pointer state	A calibrated apparatus produces one redundantly accessible form.	Consensus Armour is gone. It becomes NORMAL against Lethal Damage.
6: Objective body	The common form has been independently copied and confirmed.	It cannot change basis or leave while the record network survives.

Raise the track when the group establishes a new common invariant through a reproducible process: synchronizing accounts to fix one event time and place, independently measuring impulse or heat or sound or displacement, observing through the same calibrated instrument, predicting an action before it happens and recording the hit, broadcasting one instrument’s output to several observers live, or repeating the result with independent equipment. Plain agreement is not enough; people can talk each other into the same words without changing how their physical systems couple to the thing. Consensus falls when shared records are destroyed, when instruments lose calibration, when observers go back to incompatible direct measurement, or when the entity kills the one person who understood how several data streams line up. At Rating 6, deleting one file does nothing, because the common form is redundantly encoded; the group has to dismantle the record network or release the entity some other way.

Cthulhu Eternal statistics

Roll each manifestation or use the typical column; it is one angular organism forced into curved space, and its projected magnitudes vary.

THE DECOHERENT, one event without one body

Characteristic	Roll	Typical
STR	4D6+4	18
CON	4D6+4	18
DEX	4D6+2	16
INT	3D6+4	14
POW	4D6+4	18
HP	(STR+CON)÷2	18
WP	= POW	18

Size category: mechanically Medium. Apparent size varies by profile but does not change target modifiers.

Movement: 10 metres per combat turn. Each profile describes the movement differently.

Armour: 6 points of Consensus Armour at Consensus 0–2; 3 points at 3–4; none at 5–6.

vs Lethal Damage: RESILIENT at Consensus 0–4. A successful Lethality roll does 2D10 HP damage minus Consensus Armour rather than killing it outright. At Consensus 5–6 it is NORMAL and takes Lethal Damage like any ordinary physical organism.

Attacks:

Invariant Contact 60%, damage 1D10. The target is thrown up to 5 metres and must pass an Athletics or DEX test to stay standing. Every observer describes the attack by their own profile; the damage and displacement are rolled once.

Sever Correlation 50%, opposed by POW. On a success the target loses 1D6 WP and cannot perceive one other observer for 1D4 turns. They still see that person’s physical effects, doors opening, guns firing, objects moving, but their mind edits out the person producing them. It isolates records and can wreck a coordinated plan.

Skills: Alertness 70%, Athletics 60%, Stealth 50%, Unnatural 50%.

SAN Loss: 1/1D6 on first seeing your projection. Comparing two irreconcilable but demonstrably authentic records costs 1/1D4. Watching the projection change after you have entered a shared pointer state costs 1D4/1D8.

Observer-Relative Morphology

Each observer perceives and records it through their assigned profile, and the GM describes the same action separately where it matters. It does not get multiple actions, bodies, or HP pools. An observer can interact with anatomy that exists only in their profile: Cho can jam a muzzle against the Black Animal’s jaw even though nobody else sees a jaw, and if she succeeds you resolve damage normally while every other observer watches an equivalent injury open in their own projection. No profile is the real one.

Cross-Basis Injury

Whenever it takes 5 or more HP from one attack, every observer sees a major wound at the same instant. Ask each player to describe what would produce the rolled result in their projection, then keep those descriptions for later scenes. A recorded cross-basis injury is a shared invariant, and it raises Consensus by 1 the first time it is measured or deliberately compared.

Basis Slip

At Consensus 0–2 the Decoherent may spend 2 WP after an attack lands on it. The attacker rerolls and takes the worse result. Their projection did not dodge; the attack simply coupled badly to the underlying angular process. At Consensus 3–4 this costs 4 WP. At 5 or higher it is unavailable.

Attack the Apparatus

It knows which systems are forcing its projections toward agreement, and it goes for synchronized clocks, shared displays, calibration targets, network switches, scientific instruments, and the people translating between incompatible records. Once per scene it can identify the single local object or person contributing most to Consensus. Not omniscience, just a sense of which coupling is squeezing its state hardest.

Consensus Shock

Whenever Consensus rises, it loses 1D4 WP, and at Ratings 3, 5, and 6 it also loses 1D4 HP ignoring Armour. Every observer sees the reaction differently and all of them agree it suffered. The cry comes from several incompatible anatomies being forced to share one nervous system.

Objective Captivity

At Consensus 6 it has one stable projected body, fixed by whatever apparatus or process established the pointer state, and everyone can now see and record the same organism. It cannot escape into angular time while the redundant common record survives. Kill that body and you kill the finite projection; whether the underlying angular organism dies, goes free, or stays trapped as inert information, nobody knows. Destroying enough shared records to drop Consensus below 6 restores the observer-relative forms, and that may be the only humane option on the table.

Using private descriptions at the table

The trick is to get the characters disagreeing without turning the session into paperwork. Before you play, make one short card per Observation Profile: what the thing looks and sounds like, how Invariant Contact appears, what evidence the observer expects it to leave, their basis suppressor, and one instruction, describe this as objective fact and do not reconcile it for the other players. At first contact, hand each player their card and narrate only the invariant event aloud, location, timing, victim movement, damage, and let the players supply the morphology when they speak or act. Do not drown the game in secret notes. Once the contradiction is out, ordinary table argument is exactly the point; let them fight about it, test it, and design measurements together. If two Protagonists fire on the same turn, resolve both against the same target even though one is aiming at a skull and the other at a standing fracture. If one restrains an apparent limb, another can exploit the restraint without seeing the same limb; just describe their own projection getting constrained in an equivalent way. The mechanics stay shared. Only the fiction is observer-relative.

Cameras do not settle the argument

A recording device is just another physical observer, and it develops its own stable projection from its own coupling. Everyone watching the recording can agree about what the camera recorded, which may raise Consensus by fixing timing, location, and conserved effects, but it does not erase anybody’s direct memory or prove the camera’s morphology is the real one. Different instruments can produce wildly different creatures. Visible light records articulated cable. Infrared records a crouching human-shaped heat deficit. Lidar maps a moving cavity. Ultrasound finds a layered membrane. A pressure floor records four massive feet. A dog attacks a scent trail no human instrument can detect. The useful question was never “which device sees the truth.” It is “which devices can be made to record the same observable.”

Forced consensus

The Protagonists can deliberately build a common pointer state. They need one measured invariant that every projection affects (position, momentum transfer, emitted heat, pressure), one calibrated apparatus that maps that invariant to a common output, redundant access so several observers get the same output without independently measuring the thing through incompatible channels, a prediction the shared model gets right at least once, and preservation, so the result lasts long enough to constrain the next interaction. A crude version might put synchronized pressure sensors under an opaque enclosure and have every Protagonist watch the same derived position marker instead of looking inside. A sophisticated lab might fuse lidar, thermal, and acoustic data into one constrained estimate and distribute only that estimate. The catch is that succeeding gives the entity a stronger body. Its Armour drops because attacks now couple cleanly, but its Invariant Contact also becomes visible and comprehensible to everyone. At Consensus 5 it can be shot like an animal. It can also bite everyone with the same mouth. Truth is not automatically safe.

Its behaviour

The Decoherent is another wounded angular animal, but its injury is not a Hound’s simple forced localization. A Hound is pinned in curved space by the records of an encounter. This thing is pinned in several incompatible versions of curved space at once. Every new observer adds a body, every comparison forces those bodies to overlap, and Consensus reduces the contradiction by crushing the organism toward one finite state. So it acts to cut down observational incompatibility: it breaks the systems that make shared records, separates observers who are comparing projections, disables the most constraining sensory or measurement channel, flees regions saturated with environmental records, and attacks a living observer only when that observer is sustaining or coordinating the coupling. It might kill the lights without going near anyone. It might smash a monitor and ignore the armed person beside it. It might drag one witness behind a wall, not to eat them, but to keep their account from being synchronized with the rest. At Consensus 0 it is confused but relatively unconstrained. At 3 it is in obvious pain. At 5 it is desperate. At 6 it has been forced into one body and will destroy anything that can keep that body objective.

Failure modes

Majority rule. The group takes the description with the most witnesses behind it and writes off the dissenters as deluded. The similar profiles become redundant records, Consensus climbs, and the excluded projections stay physically real, so the consensus body forms with anatomy nobody modelled correctly.

The best camera. Investigators trust the highest-resolution recording. Resolution is not observer independence. They build a trap for the camera’s creature and find out that the pressure floor selected a different means of movement.

Forced testimony. Witnesses get coached until their language matches. The words converge; the physical couplings do not; Consensus does not rise; and under stress the incompatible memories come back.

Sensor fusion. Software blends incompatible inputs by averaging them into a model that is smooth, stable, and wrong, and if enough systems then act on it, it becomes a new pointer state anyway, forcing the Decoherent into anatomy optimized for nobody’s world.

Killing the dissenters. Someone works out that fewer observer classes mean less incompatibility and starts eliminating witnesses whose profiles differ from the dominant one. It raises Consensus efficiently. It also makes the murderer the architect of the final body.

Scenario operations

The Depositions

Four survivors of an industrial disaster are suing the manufacturer, and their mutually incompatible accounts have wrecked the case. Then discovery turns up instrument logs proving each account correctly predicted different physical evidence. The company’s lawyers hire the Protagonists to find the fraud. The Decoherent is still in the evidence warehouse, where every expert examination hands it another body.

Peer Review

Three laboratories publish irreconcilable images of the same captured organism, and each result replicates perfectly inside its own apparatus. A joint replication effort will either resolve the controversy or supply exactly the redundancy needed to imprison the thing for good.

The Hostage Camera

A tactical team goes into a building steered entirely by body-camera feeds. The operators outside see one cable-like threat. Every officer inside sees something else. Command keeps issuing clean instructions for the camera’s projection, quietly raising Consensus around a body the entry team cannot perceive.

One Witness Too Many

A remote research station has stayed stable by allowing exactly one person to observe the containment chamber at a time. A medical emergency forces a second worker inside. The two observers disagree. The automated safety systems are a third profile.

The Humane Release

The Protagonists inherit a facility holding a fully objective Decoherent, finite and visible and slowly dying. Releasing it means destroying decades of unique research and deliberately restoring incompatible observation while standing next to the thing as its single body comes apart.

Field classification

Type: Observer-relative angular organism. Habitat: Places with several incompatible measurement channels but little redundant environmental recording. Diet: None identified. Behaviour: Disrupts shared records and separates observers to reduce incompatible localization. Primary sign: Witness accounts stay individually consistent, experimentally supported, and mutually impossible. Reproduction: Unknown. The apparent multiplicity is projection, not offspring. Containment: Establish one calibrated pointer state and distribute it redundantly. Humane release: Destroy the shared pointer state and let the projections decohere from one another. Common fatal error: Assuming the most detailed observation is the most real.

Addendum to Incident 6-14

Cho brought up the pumping station’s inspection array at 21:24. Six pressure sensors, three laser gates, and the turbine’s vibration monitors all fed one diagnostic display. The software threw away the images and rendered only a white point for estimated position, and every surviving witness watched the same point cross the screen. For the first time their descriptions agreed. The black animal shed its extra legs. The burned man turned right way in inside his own skin. The standing fracture closed around a spine. The insect sheet folded itself into volume. The cables pulled taut around organs none of them had seen. One body stood in the turbine hall. It looked at all of them with the same eyes. Then it started breaking the sensors.

References and rules basis

W. H. Zurek, “Quantum Darwinism,” Nature Physics 5 (2009), 181–188: doi:10.1038/nphys1202.
M. Schlosshauer, “Decoherence, the measurement problem, and interpretations of quantum mechanics,” Reviews of Modern Physics 76 (2005), 1267–1305: doi:10.1103/RevModPhys.76.1267.
Cthulhu Eternal, “Open Rulesets,” for the common game structure: cthulhueternal.com/open-rulesets.
Cthulhu Eternal, “Open Mythos: Entities,” for creature-profile, Armour, SAN, and Lethality conventions: Entities A–K and Entities L–Z.

trollboy

I’m trollboy. Howdy.

The Decoherent