eduKateSG · VOYNICH RESEARCH LIBRARY · TANGENTIAL VOYNICH XV · FOURTEENTH FALSE WORLD
Tangential Voynich | Put the Manuscript on a Board-Game Table
A state is not defined only by what it contains. It is also defined by what it still allows.
← Tangential Voynich: The Wrong-System Protocol · Previous False World: Choreography · Voynich Research Library
Choreography gave Tangential Voynich feasibility topology.
It asked whether a transformation that looks valid on paper is actually reachable from a particular prior state, through a plausible route, under hidden constraints, and at a measurable cost.
A board game changes the question again.
A game state is not merely a picture of pieces on a board. It is a position with an option set. Some moves are legal. Some are illegal. Some are technically legal but strategically ruinous. Some moves preserve future options. Others narrow them. Some choices are reversible. Others commit the player irreversibly. A position can look rich while being almost forced. Another can look simple while containing many latent futures. A local gain can destroy a global plan. A move can be strong because of what it prevents rather than what it creates.
This makes the board-game world a useful wrong system for the Voynich Manuscript.
The manuscript already gives us states, boundaries, repeated forms, constrained continuation, local alternatives, rare transitions, multi-axis populations and many absences. The board-game tangent asks a different kind of structural question:
For a given Voynich state, what options remain available next—and how does choosing one option change the future option space?
That is not a game theory of Voynich.
It is an experiment in choice structure.
FALSE-WORLD CONTRACT
Voynich Is Not a Board Game
- A glyph is not a game piece.
- A token is not a move.
- A line is not a turn.
- A paragraph is not a round.
- A page is not a board position.
- A recurring form is not a player action.
- A rare form is not a special move.
- A boundary is not a turn change.
- A diagram is not a game board.
- A scribal hand is not a player.
- Hidden information is not evidence of cipher.
- Strategic-looking structure is not evidence that the manuscript was used for play, divination, instruction or competition.
Legal move ≠ grammar rule. Player ≠ scribe. Turn ≠ line. Strategy ≠ meaning. Hidden information ≠ cipher.
Every game term is temporary. If a result cannot be rewritten without players, pieces, winning, losing or strategy, it does not return to the Voynich evidence ledger.
What the Board Game Adds: Decision Topology
Choreography asked whether a state can reach another state.
The board game asks what happens to the rest of the reachable future once one transition is chosen.
This adds an eighth hidden geometry:
decision topology.
Decision topology studies the option space attached to a state.
Let S be a structural state. Let A(S) be the set of admissible next transitions from S.
A(S) = {a₁, a₂, …, aₖ}
After choosing action a, the system enters state S′. The important quantity is not merely whether S′ was reachable. It is how the option set changes:
ΔΩ = |A(S′)| − |A(S)|
A transition can therefore expand, preserve or collapse future possibility.
For Voynich, this becomes a measurable question about continuation entropy, branching factor, reversibility and constraint accumulation.
Legal Moves: Possibility Must Be Conditioned on State
A move can be legal in one board position and illegal in another.
This is stronger than simple bigram legality.
For a candidate Voynich transition A→B, test legality under increasingly rich state descriptions:
- immediate neighbour only;
- neighbour + line position;
- neighbour + boundary state;
- neighbour + hand;
- neighbour + Currier regime;
- full multi-axis local state.
If a transition’s apparent legality changes sharply when one state variable is added, the simpler grammar was incomplete.
admissibility belongs to a state, not to a pair of symbols in isolation.
Branching Factor: Some States Offer Many Futures, Others Almost None
A board position with twenty legal moves is structurally different from one with only one.
The Voynich analogue is local branching factor.
For every state class S, estimate the effective number of plausible continuations after frequency and opportunity correction.
Possible measures include:
- number of observed continuation classes;
- Shannon entropy of continuation;
- effective support size exp(H);
- probability mass concentrated in top-k alternatives.
A state can therefore be classified as open, moderately constrained or near-forced without assigning semantic meaning.
This gives a new way to compare line starts, middles, endings, labels and special page interfaces.
Forced Moves: Low Entropy Can Reveal a Constraint
In a game, a forced move is not necessarily the most common move globally. It is the move required in a particular position.
This distinction matters for Voynich.
A form may be rare in the corpus yet nearly mandatory after one specific state.
Search for contexts where:
H(next | S) ≪ H(next)
Then test whether the low-entropy continuation survives held-out material and state-matched controls.
The neutral residue is a near-deterministic conditional transition.
It is not proof of syntax, semantics or procedure.
Choice Collapse: One Transition Can Destroy Many Future Options
Some game moves commit the position. After the move, many alternatives vanish.
For Voynich, measure whether particular transitions sharply reduce future branching factor over the next k positions.
Define an option-retention curve:
R(k) = expected reachable continuation diversity k steps after transition a.
Compare R(k) after different transition classes.
A transition that consistently narrows future structure may be a commitment-like event.
This extends Choreography’s transition cost by asking not only what the transition costs now, but what possibility it removes later.
Option Preservation: Some Transitions Keep the System Flexible
The reverse also matters.
Some game moves preserve many possible futures. They defer commitment.
For Voynich, identify transition classes after which continuation diversity remains unusually high.
Such forms may function as flexible bridges or neutral continuations in a structural sense.
This can be tested by comparing expected future entropy after each class while controlling for position and manuscript population.
some structures may be important because they preserve optionality rather than because they dominate frequency.
Decision Depth: The Consequence of a Choice May Appear Much Later
In strategy games, a move can matter several turns later.
The local board position may not reveal the consequence immediately.
Voynich already has long-range structure questions. The board-game tangent sharpens them into decision depth.
For a transition class a at position t, measure how far into the future a improves prediction relative to a model that forgets it.
The effect horizon can be summarised as:
d* = maximum k for which I(aₜ ; Sₜ₊ₖ | local controls) remains reliably above null.
A high d* state has long structural consequence.
A low d* state is locally important but quickly forgotten.
State Value Without Winning
Games assign value to positions because some are better for achieving an objective.
Voynich has no known win condition, so “good” and “bad” states are forbidden concepts.
But we can define neutral structural value functions.
For example, a state may be high-value for prediction because it:
- strongly predicts the next boundary;
- identifies a page population;
- preserves many future options;
- acts as a bridge between regimes;
- predicts a long-range return;
- reduces uncertainty across several axes.
Call this informational state value, not strategic value.
a state can be structurally valuable because of how much future uncertainty it resolves.
Tactics and Strategy: Local Prediction and Global Structure Are Different
A tactic exploits a local situation. A strategy shapes long-range possibilities.
The Voynich analogue is scale separation.
A local rule may predict the next token well while explaining nothing about paragraph, page or bifolio structure.
A manuscript-wide model may explain large populations while predicting local continuation poorly.
Do not force one model to own both scales.
Evaluate separately:
- one-step prediction;
- boundary prediction;
- local regime persistence;
- page-class prediction;
- bifolio-level coherence;
- whole-manuscript population recovery.
local predictive success should not silently become a global theory.
Hidden Information: Unobserved State Is Not Automatically Encryption
Many board games contain information unavailable to all players at all times.
The manuscript also contains missing state, but that does not imply deliberate concealment.
Possible hidden variables include:
- lost original page order;
- unknown segmentation;
- unseen production sequence;
- missing exemplar context;
- receiver-supplied conventions;
- damaged or lost leaves;
- unobserved material features in scans.
The board-game tangent introduces a disciplined split:
- hidden by design;
- hidden by loss;
- hidden by representation;
- hidden by current measurement limits;
- not actually hidden—only not yet modelled.
unobserved ≠ secret.
Information Sets: Different Histories Can Look the Same Now
In games with hidden information, a player may be unable to distinguish several true underlying states because they produce the same visible observation.
The Water Network called this history-non-identifiable state. The Board Game formalises it as an information set.
For Voynich, group latent histories that remain observationally indistinguishable under current evidence.
Example:
- two candidate page orders consistent with the same surviving bifolia;
- two segmentation schemes producing the same local transition statistics;
- two production trajectories yielding the same final glyph outline;
- two source traditions yielding the same transformed visual component.
The correct output is not to choose one because it feels more elegant.
when evidence cannot discriminate histories, preserve the equivalence class of possible histories.
Asymmetric Roles: One Rule System Can Contain Different Participant Classes
Some games give different players different legal actions or objectives.
The Voynich analogue is role asymmetry across structural populations.
Labels, running text, paragraph initials, marginal markers and diagram text need not obey the same continuation rules.
Instead of treating deviations from one universal grammar as noise, test role-specific rule sets.
A stronger model might be:
shared global constraints + role-conditioned local move sets.
If role-conditioned models improve held-out prediction without proliferating arbitrary exceptions, structural asymmetry gains weight.
Turns: Alternation Must Be Demonstrated, Not Imagined
Board games often alternate turns. That makes “turn-taking” a seductive but dangerous metaphor.
Voynich has multiple hands and multiple structural classes, but alternating patterns do not automatically represent agents taking turns.
If alternation is proposed, test for:
- stable A/B/A/B sequencing beyond chance;
- role-specific allowed transitions;
- consistent handoff markers;
- resets when the alternation breaks;
- persistence across pages or sections.
Without those properties, “turn” should disappear from the model.
The neutral residue is structured alternation.
Tempo and Initiative: Who Changes the State First?
In many games, one side can force the other to respond. The value lies in controlling the next decision.
The Voynich analogue is directional constraint propagation.
For a candidate marker or boundary class, ask whether it:
- reduces the continuation options of the next unit;
- predicts a specific response family;
- changes several features before they can independently vary;
- creates a stable lead-lag order across axes.
The Choreography tangent called this cue propagation.
The Board Game adds option suppression: a state may exert influence because it constrains what the next state is allowed to choose.
Traps: Locally Legal Paths Can Lead to Structurally Dead Futures
A trap is not an illegal move. It is a legal move whose downstream consequences are unexpectedly restrictive.
For Voynich, search for transitions that appear locally ordinary but are followed by:
- very low continuation entropy;
- high probability of immediate closure;
- restricted terminal families;
- reduced cross-regime connectivity;
- unusually high correction rate where corrections are physically verified.
The game story disappears. The neutral object is a locally admissible transition with strongly narrowing downstream state space.
This is a different class from simple forbidden transition.
Dead Positions: A State Can Be Stable but Have No Productive Continuation
Some game positions are effectively dead: the outcome is fixed or no meaningful progress remains possible.
Voynich cannot inherit the game’s outcome semantics, but it can be tested for absorbing structural states.
Candidate dead-state properties include:
- very restricted continuation;
- high terminal probability;
- no return to ordinary running-text regime;
- repeated self-looping without structural expansion.
Water Network called these absorbing endpoints. Board Game adds a future-option interpretation.
a state can be non-terminal in appearance yet terminal in option space.
Endgames: Rules Can Change When the Remaining State Space Shrinks
Many games behave differently near the end because fewer pieces, spaces or options remain.
The Voynich analogue is terminal-regime change.
Line endings, paragraph endings and page endings may not simply be ordinary sequence continuation closer to a boundary. They may enter a distinct rule regime.
Compare models:
- one stationary grammar across entire line;
- smooth positional gradient;
- discrete terminal regime beginning at learned distance from end.
If a discrete terminal regime predicts better on held-out lines, “endgame” language can be removed and replaced by boundary-proximal rule shift.
Opening Theory Without Openings
Games often contain stereotyped opening families because early choices establish later structure.
Voynich already shows strong line-start and paragraph-start effects.
The board-game tangent asks whether launch states should be evaluated by the future option spaces they create.
For each launch family, compare:
- future branching factor;
- probability of later family classes;
- boundary distance;
- long-range return signatures;
- cross-regime persistence.
A line-start form may be important not because it “means beginning,” but because it selects a continuation regime.
launch states can be classified by the future spaces they make reachable.
Reversible and Irreversible Choices
Some game moves can effectively be undone. Others permanently alter the position.
Choreography introduced graded irreversibility. Board Game adds future-state asymmetry.
For transition A→B, compare:
- probability of returning to A-like state;
- number of steps required;
- whether the return restores the original option set;
- whether the apparent return is only surface similarity.
A true reversible transition should restore not just appearance but comparable future reachability.
return to the same-looking state is not full reversal unless the option space also returns.
Sacrifice: Local Loss Can Preserve a Larger Structure
In games, a player may give up a local asset to preserve or improve a larger position.
The literal strategic meaning is unusable for Voynich.
The neutral question is whether local simplification supports higher-order constraint.
Under spatial pressure, does the manuscript drop optional detail while preserving:
- core family identity;
- boundary role;
- page-level distribution;
- cross-layer alignment;
- future transition regime?
Power Grid called this load shedding. Choreography called it density-conditioned realisation. Board Game reframes the residue as local simplification that protects higher-order structure.
Dominated Options: Some Alternatives May Be Structurally Redundant
A dominated move is an option that is never preferable under a defined objective because another option achieves at least as much in every relevant dimension.
Voynich has no known objective, so dominance cannot be semantic.
But we can test representational redundancy.
If two candidate forms occur in the same contexts and produce indistinguishable downstream state distributions, one may add little structural information beyond the other.
Quantify whether the distinction improves prediction.
If not, the two forms may belong to one broader behavioural class despite surface difference.
a distinction that never changes admissible futures may be representationally finer than the system requires.
State Compression: Different Positions Can Have the Same Future
Game analysis often treats positions as equivalent when they lead to the same relevant future structure despite superficial differences.
This gives Voynich another route to equivalence.
Instead of clustering states by surface form, cluster them by future transition distribution.
Two states S₁ and S₂ can be considered future-equivalent when:
P(future | S₁) ≈ P(future | S₂)
under predefined distance and scope.
This is potentially powerful because it lets the system define equivalence through consequence rather than appearance.
Music asked what transformations preserve identity. Board Game asks whether identity can be defined by preserved future option space.
Transpositions of State: Different Surfaces, Same Decision Structure
Two board positions may be structurally equivalent under symmetry even if piece coordinates differ.
For Voynich, test whether two local structures have equivalent continuation graphs after a frozen relabelling or positional normalisation.
Examples:
- same continuation topology near line start and line end after coordinate normalisation;
- same label-family transition structure across different diagram sectors;
- same hand-conditioned option graph under surface substitution.
This is stricter than ordinary pattern similarity because the entire reachable-future structure must map coherently.
The neutral object is a state-graph isomorphism candidate.
Rule Exceptions: Special Moves Must Pay Rent
Board games often contain special rules, but every special rule is explicitly defined.
Voynich theories often accumulate exceptions more casually.
The Board-Game tangent imposes an exception budget.
Every proposed special-case rule must specify:
- the exact triggering state;
- the exceptional transition allowed or forbidden;
- why the ordinary rule fails there;
- how often the trigger occurs;
- whether the rule predicts held-out cases.
A model that needs a new exception for every inconvenient page is not learning a rule system.
special rules must be rarer than the evidence they explain.
Rule Discovery: Infer the Move Set Before Inferring the Objective
If you discover an unknown game without its rulebook, the first task is not to guess who wins.
You first infer what moves are legal, what states exist, what transitions recur and where the system terminates.
That sequence is directly useful for Voynich.
Before asking what the manuscript says or does, infer:
- candidate units;
- state variables;
- transition constraints;
- boundary changes;
- role-specific move sets;
- reachable-state structure;
- termination and reset conditions.
Only after those survive should higher-order purpose hypotheses be allowed to compete.
infer the rule space before inferring the goal.
Search Trees: A Model Should Expose Its Alternative Futures
A board-game engine does not evaluate only the move it chooses. It explores alternative continuations.
Tangential Voynich should do the same.
For any proposed interpretation of a boundary, token family or page transition, retain the competing continuation tree.
Do not report only the path that happened to fit.
At each step, store:
- candidate alternatives;
- their predicted likelihood;
- which observation removed each alternative;
- which alternatives remain observationally equivalent.
This creates a defeat ledger for paths not taken.
a conclusion is stronger when we can show which nearby alternatives the evidence actually eliminated.
Pruning: Eliminate Futures Only With Evidence
Search becomes efficient when impossible or dominated branches are pruned.
Research has the same temptation, but premature pruning is dangerous.
A Voynich hypothesis branch should be removed only when a stated observation conflicts with it or when a stronger model subsumes it under a preregistered criterion.
Do not prune because a branch is unfashionable, aesthetically awkward or historically inconvenient.
Preserve branch status:
- active;
- weak;
- dominated;
- contradicted;
- unobservable;
- not yet tested.
This maps naturally onto the existing evidence ladder and defeat ledger.
Lookahead and Overfitting: More Depth Can Make a Bad Model Look Smarter
A search that explores more future branches can find impressive lines of play. A research model with enough flexibility can also find impressive post-hoc explanations.
Deeper modelling therefore needs stronger held-out discipline.
As decision depth increases, track:
- number of free parameters;
- number of allowed exceptions;
- training fit;
- held-out fit;
- null-control fit.
If training fit improves while held-out discrimination collapses, the search is memorising the manuscript.
deeper foresight is useful only when future prediction improves faster than model freedom.
Local Minima: A Stable Interpretation Can Still Be Wrong
Players and algorithms can become trapped in locally attractive positions that block better long-range outcomes.
Voynich research can become trapped in locally coherent interpretations.
A plant identification may explain one page beautifully while forcing awkward interpretations elsewhere.
A cipher mapping may fit selected strings while destroying line-position effects.
A provenance story may explain an intellectual environment while lacking object-level evidence.
So every locally strong model should be challenged by a wider scorecard.
local coherence is not global optimality—and neither is proof.
Zugzwang Without Zugzwang: When Every Available Transition Narrows the Future
Some game states are difficult because every available move worsens the position.
The strategic meaning does not transfer, but the option-space geometry does.
Search for Voynich states where every observed continuation reduces future branching factor or forces near-term closure.
If such states exist, they represent unavoidable structural commitment points.
The neutral return is:
states from which all available continuations reduce future optionality.
This can help distinguish ordinary narrowing from structurally inevitable narrowing.
Repetition and Draw Loops: A Cycle Can Preserve State Without Progress
Games can cycle through repeated positions.
The Water Network already asked about recirculation. Board Game adds state equivalence.
A recurring sequence should be tested for whether it returns to:
- the same surface state;
- the same behavioural class;
- the same option set;
- the same multi-axis context.
Only the last two justify calling the loop a state return rather than mere repetition.
This strengthens Music’s long-range return test.
Chance and Determinism: Variation Can Enter at Different Layers
Some games are deterministic; others include dice, cards or shuffled hidden states.
Voynich variation may likewise enter at different levels.
Candidate models include:
- deterministic transition rules with noisy transcription;
- probabilistic choice among legal families;
- hand-specific stochastic realisation;
- mixture of local generators;
- copy-modify process with constrained random variation.
The board-game tangent warns against forcing all uncertainty into one layer.
randomness in observation, choice, production and model error are different things.
Perfect Information and Imperfect Information Are Different Controls
A chess-like control exposes every board state. A card-game-like control hides some state.
Reverse Tangential calibration should include both.
If a method succeeds only when all latent state is visible, then applying it to an incomplete manuscript may generate false certainty.
Therefore calibrate separately on:
- fully observed deterministic games;
- fully observed probabilistic games;
- partially observed games;
- games with asymmetric information.
The aim is not to identify Voynich with any game class. It is to learn which structural features remain recoverable under missing state.
The Board Itself: Geometry Constrains Action
A game piece’s legal moves often depend on where it sits on the board.
This reinforces a central Voynich principle: geometry is not decoration if geometry changes the admissible action set.
For lines, labels and diagrams, test whether local position predicts:
- available token-family classes;
- component size;
- transition entropy;
- boundary probability;
- recurrence structure.
Choreography called this page-zone conditioning. Board Game states it as a rule-space problem:
the geometry of the surface may determine which structural actions are admissible.
Territory Without Territory: Occupied Space Can Change Future Availability
Board games often make occupancy matter. A square occupied now can block later movement.
For Voynich, the neutral analogue is representational occupancy.
Text, labels, images and diagram components share finite page space. Once one element occupies a region, other options disappear or become compressed.
Model how occupied area changes:
- label count;
- token length;
- spacing;
- route of surrounding text;
- density of visual components.
This extends Choreography’s collision avoidance into a dynamic future-space model.
What has already been placed changes what can still be placed.
Local Rules, Global Rules and Meta-Rules
Games can have rules that apply to individual pieces, rules that apply to the whole position, and special rules governing when ordinary rules change.
The Voynich model may need similar layering.
- local rules: immediate token or component transitions;
- regional rules: line, paragraph, page or interface constraints;
- population rules: hand or Currier-specific parameters;
- meta-rules: conditions under which one regime changes into another.
The important discipline is to test whether each additional layer earns predictive gain.
A layered model should not become a licence for unlimited exceptions.
Its complexity must be paid for in held-out performance.
Decision Boundaries: Some Marks May Change the Available Move Set
Compiler, Operating System, Water Network and Choreography have all produced boundary effects.
Board Game asks a particularly strong version:
does crossing this boundary change the set of admissible future states?
Measure the option set immediately before and after candidate boundaries.
Then compare matched non-boundary positions.
A boundary that merely changes frequency is one thing.
A boundary that removes entire continuation classes is stronger.
This may become one of the cleanest cross-world invariant measurements in the programme.
Decision Topology and the Missing Receiver
Games are only games because someone knows which options are meaningful.
The board alone may not reveal the objective, the value of pieces, the consequences of hidden information or why one legal move is preferable.
This is relevant to Voynich because the original receiver is missing.
We may be able to reconstruct option sets without reconstructing preference.
We may recover legal-looking transition structure without knowing why a trained user would choose one branch over another.
choice architecture can survive after the preference function is lost.
This is a particularly important separation for the broader Tangential representation branch.
The Decision Ledger
| State | Neutral definition |
|---|---|
| Open | Many continuation classes remain admissible. |
| Constrained | Continuation diversity is substantially below matched baseline. |
| Forced | One continuation family dominates under a specific state. |
| Commitment | A transition sharply reduces downstream option diversity. |
| Flexible | A transition preserves unusually broad future option space. |
| Prepared | A continuation becomes admissible only after a specific prior state. |
| Trap-like | A locally admissible transition leads to severe downstream narrowing. |
| Absorbing | The system rarely or never returns to ordinary continuation after entry. |
| History-ambiguous | Several latent paths produce the same visible state. |
| Future-equivalent | Distinct states induce approximately the same continuation distribution. |
| Irreversible | Return to the prior option space is improbable or impossible. |
| Unobservable | Available evidence cannot distinguish important latent states. |
None of these labels implies game semantics.
They are candidate states for a structural transition system.
Reverse Tangential Test: Hide a Real Game
The board-game tangent earns credibility only if Voynich-style methods can recover rule and decision structure in known games after ordinary labels are removed.
Use several control classes rather than one favourite game.
- a deterministic perfect-information game;
- a stochastic perfect-information game;
- a hidden-information game;
- an asymmetric-role game.
Remove piece names, player names, move names, objectives and strategic annotations.
Preserve neutral state-transition data.
Can the methods recover:
- legal versus impossible transitions;
- branching factor;
- forced states;
- commitment moves;
- reversible versus irreversible transitions;
- role asymmetry;
- terminal regimes;
- state equivalence by future options;
- hidden-state ambiguity;
- option-preserving versus option-destroying transitions?
Then damage the controls.
Delete state variables. Merge piece classes. Hide move order. Introduce observation noise. Preserve marginal transition counts while shuffling state context.
If the method still “discovers” rich decision structure in controls whose rule structure has been destroyed, it is manufacturing a game.
A method that cannot recover hidden move constraints in a known rule system should not invent decision topology in Voynich.
What Would Make the Board-Game World Fail?
| Board-game projection | Failure condition | Neutral residue |
|---|---|---|
| Legal move set | State variables do not alter continuation admissibility beyond local frequency. | Simple transition model may suffice. |
| Forced state | Low-entropy continuations disappear on held-out data. | No stable near-deterministic context. |
| Choice collapse | Transitions do not change future option diversity beyond position controls. | No commitment effect. |
| Hidden information | Latent-state models add no explanatory value. | Observed state may be sufficient. |
| Role asymmetry | Role-conditioned models do not outperform one shared rule set. | Universal rule model remains preferable. |
| Trap-like state | Locally admissible transitions do not predict downstream narrowing. | No trap-like geometry. |
| Future-equivalent state | Surface-different states do not share stable continuation distributions. | No consequence-based equivalence. |
| Decision boundary | Candidate boundaries do not change reachable option classes. | Boundary effect may be frequency-only. |
| Decision topology | Option-space variables do not improve held-out prediction or discrimination from null controls. | Simpler sequence/feasibility models remain sufficient. |
The Board Game is allowed to lose completely.
If legal-move, branching and future-option concepts add nothing beyond existing transition statistics, the tangent collapses back into simpler models.
The Board-Game Experiment Pack
- State-conditioned legality: test transition admissibility under increasingly rich state descriptions.
- Branching-factor map: measure effective continuation diversity by position, boundary and population.
- Forced-state scan: identify contexts with sharply reduced conditional entropy.
- Choice-collapse test: measure whether transitions destroy downstream continuation diversity.
- Option-preservation test: identify transitions that maintain unusually broad future state space.
- Decision-depth analysis: estimate how long one transition continues to affect future state prediction.
- Informational state value: rank states by how much future uncertainty they resolve.
- Scale separation: evaluate local and manuscript-wide models independently.
- Hidden-state ledger: distinguish designed concealment, loss, representation limits, measurement limits and unmodelled state.
- Information-set modelling: preserve observationally indistinguishable latent histories.
- Role-conditioned rule sets: compare labels, running text, initials and interfaces under shared versus asymmetric constraints.
- Alternation test: require structured alternation before any turn-taking metaphor survives.
- Option-suppression propagation: test whether candidate states constrain what later states can choose.
- Trap scan: identify locally admissible transitions followed by severe downstream narrowing.
- Absorbing-state analysis: detect states with little return to ordinary continuation.
- Terminal-regime comparison: distinguish positional gradient from discrete end-state rule shift.
- Launch-family future map: classify starting states by the future regimes they make reachable.
- Reversibility audit: require restoration of option space, not merely visual return.
- Local-simplification test: ask whether dropping detail preserves higher-order structure under pressure.
- Representational dominance test: merge distinctions that never change downstream behaviour.
- Future-equivalence clustering: group surface-different states by continuation distribution.
- State-graph isomorphism: test relabelled or position-normalised continuation graphs.
- Exception budget: require every special-case rule to have a frozen trigger and held-out prediction.
- Rule-before-goal reconstruction: infer state, transition and termination structure before purpose hypotheses.
- Alternative-future ledger: store nearby branches eliminated by each observation.
- Evidence-based pruning: remove hypothesis branches only under stated contradiction or dominance criteria.
- Depth/complexity curve: compare lookahead gain against model freedom and held-out performance.
- Local-minimum challenge: retest strong local explanations against manuscript-wide constraints.
- Mandatory-narrowing scan: identify states where all continuations reduce future option space.
- Cycle-state test: distinguish surface repetition from return to equivalent option space.
- Uncertainty-layer decomposition: separate stochastic production, choice, observation and model error.
- Perfect/imperfect-information controls: calibrate recoverability under different levels of hidden state.
- Geometry-to-rule test: measure how page position changes admissible structural actions.
- Occupancy-future test: quantify how existing page elements reduce later representational options.
- Layered-rule model: compare local, regional, population and meta-rule levels.
- Decision-boundary test: measure whether boundaries change admissible continuation classes.
- Preference-function firewall: separate reconstructed option sets from unknown receiver preferences.
- Reverse hidden-game calibration: anonymise known games and recover rule and option structure.
- Destroyed-rule negative controls: preserve marginal counts while breaking state-conditioned legality.
- Held-out Voynich gate: freeze unseen pages before state-value and option thresholds are tuned.
- Metaphor removal: rewrite every surviving result without game vocabulary.
Fourteen Wrong Systems, Eight Hidden Geometries
The false-world estate now contains fourteen operational systems:
- Railway: movement through routes.
- Operating system: state.
- Compiler: representation.
- Database: relationship.
- Filesystem: logical versus physical organisation.
- Telecommunications: protocol versus payload.
- Warehouse: retrieval.
- Global supply chain: dependency.
- Airport control tower: coordination.
- Power grid: balance.
- Water network: path-dependent flow.
- Musical score: transformed equivalence.
- Choreography: feasibility and transition cost.
- Board game: option sets and decision topology.
The geometry stack is now:
| Topology | Question |
|---|---|
| Retrieval topology | How is distributed structure found? |
| Dependency topology | What depends on what? |
| Coordination topology | How do local sequences coexist? |
| Balance topology | What higher-order condition constrains local behaviour? |
| Flow topology | Does present state retain information about route and mixture history? |
| Equivalence topology | Which transformations preserve identity? |
| Feasibility topology | Which transformations are reachable, through what path and at what cost? |
| Decision topology | What option set does a state expose, and how does each transition change the future possibility space? |
This is a meaningful progression.
Choreography says two states may be connected by a costly but feasible path.
Board Game says the path itself may matter because taking it changes what remains possible afterwards.
Reachability describes a path. Decision topology describes the future that path leaves behind.
Metaphor Removal
Now remove the board game.
Remove players, pieces, turns, tactics, strategy, winning, losing, openings and endgames.
What survives?
- Transition admissibility should be conditioned on the full relevant state rather than symbol pairs alone.
- Different states can expose very different continuation diversity.
- Some contexts can produce near-deterministic continuation.
- One transition can sharply reduce or preserve downstream option diversity.
- Structural consequences can persist many positions beyond the originating transition.
- States can be ranked by how much future uncertainty they resolve without assigning semantic value.
- Local predictive rules and global structural rules should be evaluated separately.
- Unobserved state can arise from design, loss, representation, measurement limits or incomplete modelling.
- Several latent histories may remain observationally indistinguishable and should be preserved together.
- Different structural roles may have different local rule sets under shared global constraints.
- Alternation requires direct evidence and should not be inferred from mere population difference.
- Some states may constrain later options by suppressing available continuation classes.
- A locally admissible transition can lead to severe downstream narrowing.
- Some states may be absorbing in option space without looking visually terminal.
- Boundary-proximal regions may operate under a distinct rule regime.
- Launch states can be classified by future regimes they make reachable.
- Reversibility requires restoration of future option space, not merely surface state.
- Local simplification may preserve higher-order structure under capacity pressure.
- Surface distinctions that never affect future behaviour may be finer than the structural system requires.
- States can be grouped by future continuation distributions rather than appearance.
- State-graph equivalence can be tested under frozen relabelling or positional transformation.
- Every exception rule needs an explicit trigger and predictive payoff.
- State and transition structure should be inferred before purpose or objective.
- Alternative futures eliminated by evidence should remain in a defeat ledger.
- Hypothesis branches should be pruned by evidence, not narrative preference.
- Deeper models require stronger held-out discrimination against nulls.
- Locally coherent interpretations can still fail global constraints.
- Some states may force all available continuations toward reduced optionality.
- Repetition should be distinguished from return to equivalent future state.
- Randomness in production, choice, observation and model error should remain separate.
- Recoverability should be calibrated under both complete and incomplete state information.
- Page geometry can change the admissible structural action set.
- Existing page occupancy can constrain later representational possibilities.
- Local, regional, population and meta-rule levels can be tested separately.
- Some boundaries may reorganise the set of admissible future states.
- Option architecture can survive even when the original receiver’s preference function is lost.
The game disappears.
The option space remains.
World Return
The Voynich Manuscript is not a board game.
But the board-game tangent exposes a weakness in ordinary sequence analysis.
We often ask what comes next.
That is only the first layer.
A stronger question is what could have come next.
Then: which of those possibilities were removed by the choice that actually occurred?
Then: how long did that constraint persist?
Then: did the choice preserve flexibility, create a forced sequence, cross into a new regime, or make return structurally difficult?
This matters because the Voynich Manuscript may contain structure not visible in raw frequencies or immediate transitions.
A form can be important because it changes the future option space.
A boundary can be important because it changes what becomes legal afterwards.
A launch state can be important because it selects a whole continuation regime.
Two surface-different states can be structurally equivalent because they preserve the same future possibilities.
Choreography asked whether the manuscript can get from here to there. The Board Game asks what future becomes possible—or impossible—because it did.
If option-space modelling adds nothing on held-out pages, the board-game world loses.
If it survives, decision topology becomes another neutral object that no longer requires the metaphor that discovered it.
Tangential Voynich: The Wrong-System Protocol → · Choreography Experiment → · Voynich Research Library →
Next Tangent: treat the manuscript as a biological regulatory network. Activation, inhibition, feedback, threshold effects, coupled modules, robustness, developmental state, attractors and perturbation response will ask whether some Voynich regularities are best understood not as sequences of choices but as distributed regulation among interacting variables.
NEXT FALSE WORLD · BIOLOGICAL REGULATORY NETWORK
The Board Game asked which future possibilities remain after a transition. The regulatory-network tangent removes the imagined chooser and asks whether several distributed variables promote, suppress, stabilise or redirect one another simultaneously.
Continue to Tangential Voynich | Treat the Manuscript as a Biological Regulatory Network →