artifacts/standard-named

Rendered from markdown source. Open raw source on GitHub.

--- title: "Navigation, Deliberation, and Multiobjective Decision" subtitle: "Optimization, MCDA, Robustness, Argumentation, Social Choice, and Human–AI Deliberation" artifact_date: "2026-07-15" artifact_type: "adjacent-fields-and-decision-review" domain: "TELIC-FIELDS" scope: "WORKING" lineage: "THE-TELIC-FIELD-PAPERS" status: "draft" processing_tier: 4 source_role: "research-and-claim-boundary-artifact" content_canon_status: "unset" publication_status: "unpublished" series_position: "G.10" companion_to:

• "20260715__TELIC-FIELDS__PAPER__CANDIDATE__F-9__polytelometric-navigation.md"

research_note: > This report compares polytelometric navigation with established methods in multiobjective optimization, multicriteria decision analysis, robust decision-making, problem structuring, argumentation, negotiation, social choice, deliberation, scenario planning, and human–AI decision support. It is not a systematic review. ---

Navigation, Deliberation, and Multiobjective Decision

Executive finding

Polytelometric navigation enters a mature and highly developed decision-support landscape.

Multiobjective optimization already models several objective functions and identifies nondominated or Pareto-efficient alternatives. Multiobjective sequential decision-making already distinguishes cases in which scalarization is impossible, infeasible, or undesirable. Multicriteria decision analysis already provides structured methods for comparing alternatives, eliciting values, applying weights or outranking, examining sensitivity, and supporting judgment. Value-focused thinking already emphasizes generating alternatives from values rather than selecting only among inherited options. Robust decision-making and decision-making under deep uncertainty already explore strategies across many plausible futures, identify vulnerabilities, and favor adaptive or satisficing routes over fragile nominal optima. Problem-structuring methods already address ambiguity, conflict, and wicked problems before formal optimization. Computational argumentation already models claims, reasons, attack, support, and acceptability. Negotiation support, group decision support, social-choice theory, and deliberative systems already address multi-party decision processes and collective preference.

The Telic Field Papers should not claim to invent:

• plural-objective decision-making;
• Pareto frontiers;
• multicriteria comparison;
• robust or adaptive planning;
• value elicitation;
• problem structuring;
• argument mapping;
• negotiation;
• social-choice limits;
• participatory deliberation;
• mixed-initiative decision support.

The candidate contribution is the composition:

Polytelometric navigation operates as a standing, consent, authority, provenance, and lifecycle layer before and around established decision methods, preserving protected and unresolved fields that should not be silently converted into objectives, weights, votes, or aggregate summaries.

This contribution remains useful only if it changes decisions in measurable ways.

The strongest candidate additions are:

1. telic admission before objective formalization;
2. source and evidence status for every represented end;
3. explicit distinction among goals, preferences, boundaries, obligations, predictions, and protected conditions;
4. cost-bearer and absent-standing maps;
5. consent and authority gates;
6. unresolved remainder as a legitimate output;
7. route lifecycle, expiry, revision, and release;
8. authority reduction and stopping when the field is inadequate;
9. model non-sovereignty in route generation and summary;
10. witness linking decision rules to later consequence.

The central boundary is:

Polytelometric navigation is not a superior universal decision rule. It is a governance process for determining what any decision rule may legitimately operate upon.

---

1. Review questions

G.10 asks:

• What do existing methods already contribute to plural-end decisions?
• Which methods require scalarization, and which preserve partial order?
• How are preferences and weights elicited?
• Which methods generate routes rather than merely rank them?
• How do methods treat deep uncertainty?
• How are reasons, dissent, and minority views represented?
• What are the formal limits of preference aggregation?
• What does deliberation add beyond voting or preference collection?
• How do human–AI systems affect reliance, framing, and representation?
• What remains distinctive in polytelometric navigation?

---

2. Multiobjective optimization

Multiobjective optimization studies problems with several objective functions.

In general, no single feasible solution optimizes every objective simultaneously. Methods therefore identify:

• Pareto-efficient solutions;
• Pareto fronts;
• tradeoff surfaces;
• compromise solutions;
• solutions selected under a scalarization or preference model.

Strong overlap

• multiple objectives;
• explicit conflicts;
• dominance;
• efficient alternatives;
• tradeoff visualization;
• sensitivity;
• sequential decision;
• interactive exploration.

Core boundary

MOO begins with a formal objective set and feasible decision space.

That set may omit:

• affected centers;
• privacy;
• consent;
• authority;
• protected conditions;
• unmodeled future cost;
• refusal;
• the option to stop or reframe the problem.

Pareto efficiency is relative to the admitted objectives.

A route can be efficient in a constitutionally impoverished model.

Adjudication

Use Pareto analysis confidently after telic admission and protected-condition review.

Do not present the Pareto front as a complete representation of legitimate choice.

Important prior finding

Research on multiobjective sequential decision-making explicitly recognizes cases in which reducing the problem to one objective is impossible, infeasible, or undesirable. Polytelometry should cite this rather than implying that resistance to scalarization is new.

---

3. Decision support beyond the Pareto front

A Pareto front may contain too many alternatives for practical choice.

Decision-support research therefore develops methods for:

• visualization;
• clustering;
• representative solution selection;
• preference elicitation;
• uncertainty exploration;
• explanation;
• interactive search;
• ethical analysis.

Recent surveys explicitly frame the problem as what occurs beyond generation of the Pareto front.

Strong overlap

• human interpretation of tradeoffs;
• interactive exploration;
• explanation;
• ethics;
• uncertainty;
• selecting among efficient alternatives.

What polytelometry may add

The telic framework asks whether:

• all legitimate centers entered the frontier;
• protected conditions were modeled as constraints or mistakenly as objectives;
• weights carry authority;
• the solution set includes fork, pause, or exit;
• the represented cost bearers consented;
• summary and visualization preserve minority standing.

Adjudication

Polytelometry belongs primarily before and around the Pareto front, not as a new algorithm for generating it.

---

4. Multicriteria decision analysis

MCDA includes families of methods for structuring and evaluating alternatives across multiple criteria.

Methods vary substantially.

They include:

• multi-attribute value or utility models;
• analytic hierarchy and network processes;
• outranking methods;
• goal programming;
• reference-point and aspiration methods;
• dominance and sensitivity analysis;
• interactive methods.

Belton and Stewart emphasize MCDA as decision aid rather than a machine that objectively determines the answer.

Strong overlap

• problem structuring;
• criteria development;
• alternatives;
• weighting;
• value judgment;
• sensitivity;
• transparency;
• facilitated decision;
• qualitative and quantitative input.

Candidate distinction

Polytelometry gives stronger independent status to:

• standing;
• evidence class;
• consent;
• authority;
• protected conditions;
• unresolved remainder;
• lifecycle and release.

Ordinary MCDA can already represent some of these through careful design.

The difference is whether they are mandatory constitutional dimensions rather than optional analyst choices.

Adjudication

Polytelometric navigation should be tested as an MCDA pre-processing and governance profile.

It should not be marketed as replacing MCDA.

---

5. Weight elicitation

Weights in decision models may represent:

• importance;
• tradeoff rate;
• value difference;
• swing;
• probability;
• priority;
• voting power.

These meanings are not interchangeable.

A common source of error is asking participants to state “importance” weights and then treating them as cardinal tradeoff rates.

Telic concern

A weight may conceal:

• source ambiguity;
• unequal authority;
• unstable preference;
• scope;
• coercion;
• incomparable values;
• protected conditions;
• interpersonal comparison.

Adjudication

Every weight should identify:

meaning
source
elicitation method
authority
scope
temporal validity
uncertainty
prohibited trades

Weight provenance is a practical candidate contribution.

Boundary

Not every decision needs formal weights.

Ordinal, threshold, outranking, or deliberative approaches may be more appropriate.

---

6. Outranking and incomparability

Outranking methods allow an alternative to be considered at least as good as another without requiring one complete compensatory value function.

They can incorporate:

• preference thresholds;
• indifference thresholds;
• veto thresholds;
• incomparability.

This is highly relevant to protected conditions.

Strong overlap

• noncompensatory reasoning;
• veto;
• partial order;
• incomparability;
• thresholds;
• imperfect information.

Candidate difference

Polytelometry links veto or noncompensation to:

• source standing;
• authority;
• consent;
• protected status;
• review.

A veto threshold selected by an analyst is not equivalent to a protected boundary asserted by an affected center.

Adjudication

Outranking methods are a strong implementation bridge.

The framework should not imply that weighted sums are the only MCDA architecture.

---

7. Value-focused thinking

Value-focused thinking begins with values and uses them to generate objectives and alternatives, rather than accepting a fixed menu of options and comparing them.

Strong overlap

• value elicitation;
• alternative generation;
• objectives hierarchy;
• consequences;
• creative routes;
• strategic thinking;
• decision opportunities.

What polytelometry may add

• multiple centers of standing;
• source and inference status;
• consent;
• authority;
• boundaries;
• unresolved conflict;
• temporal and outer-loop fields.

Adjudication

Route generation from telic fields should explicitly acknowledge value-focused thinking.

The candidate distinction is standing-aware and consent-aware generation across several centers.

---

8. Problem structuring and wicked problems

Problem-structuring methods address situations where:

• goals are disputed;
• boundaries are unclear;
• stakeholders disagree;
• causality is uncertain;
• the problem changes during inquiry;
• no final formulation is neutral.

Rittel and Webber's account of wicked problems is a foundational statement that many planning problems cannot be solved through straightforward technical optimization because problem definition and solution are entangled.

Strong overlap

• framing;
• stakeholder perspectives;
• contested goals;
• no final formulation;
• iterative learning;
• uncertainty;
• social complexity.

Candidate distinction

Polytelometry adds:

• telic projection records;
• field classes;
• authority and consent gates;
• route lifecycle;
• semantic witness;
• standing-aware stopping.

Adjudication

Polytelometric navigation is closest to problem structuring plus MCDA plus consent governance.

This is the strongest honest positioning.

---

9. Robust decision-making under deep uncertainty

Robust decision-making and related DMDU methods reject a simple predict-then-act approach where probabilities or models are deeply uncertain or disputed.

They often use:

• many plausible futures;
• exploratory modeling;
• scenario discovery;
• adaptive strategies;
• signposts;
• satisficing;
• regret;
• vulnerability analysis;
• stakeholder deliberation.

Strong overlap

• scenario navigation;
• robustness;
• adaptive route;
• uncertainty;
• review trigger;
• no single forecast;
• plural assumptions.

Candidate addition

Polytelometry asks:

• robust for which center;
• which protected conditions define failure;
• who selected the robustness metric;
• whose cost is repeated across scenarios;
• what consent governs adaptive changes;
• when a robust institutional strategy is fragile for a minority.

Adjudication

Robust decision-making is the strongest neighbor for route testing under uncertainty.

Polytelometry should extend its performance measures with standing and authority rather than invent a parallel robustness method.

---

10. Scenario planning

Scenario planning creates structured descriptions of plausible futures to test strategies, assumptions, and vulnerabilities.

Scenarios are not predictions.

They support:

• strategic learning;
• contingency;
• challenge to assumptions;
• option generation;
• adaptive planning.

Strong overlap

• temporal standing;
• future possibility;
• route stress testing;
• counterfactual divergence;
• adaptive review.

Telic risk

Scenario sets can be curated to favor one route.

Low-probability high-consequence conditions may be excluded.

Affected communities may be represented through institutional stereotypes.

Adjudication

Scenario provenance should be explicit.

Record:

• who proposed the scenario;
• which assumptions it challenges;
• which centers appear;
• which are absent;
• what would falsify it;
• what action it is being used to justify.

---

11. Argumentation frameworks

Dung's abstract argumentation framework models arguments and attack relations and defines semantics for acceptable sets of arguments.

Later computational-argumentation research adds support, preferences, values, dialogue, burden, and explanation.

Strong overlap

• disagreement;
• reasons;
• attack;
• support;
• defensibility;
• contestability;
• formal semantics;
• dialogue.

Candidate difference

An argumentation framework may represent a strong argument without representing:

• source standing;
• bodily or low-verbal evidence;
• consent;
• cost bearer;
• authority;
• temporal validity;
• protected condition.

Adjudication

Use argumentation as the reason layer within the navigation record.

Do not substitute argumentative acceptability for legitimate route authority.

---

12. Negotiation support

Negotiation-support systems assist parties in:

• preference elicitation;
• offer generation;
• tradeoff discovery;
• concession;
• mediation;
• agreement drafting;
• outcome comparison.

Strong overlap

• different centers;
• conflicting ends;
• route generation;
• compromise;
• reservation points;
• agreement;
• mediation.

Candidate difference

Polytelometry distinguishes:

negotiable preference
protected condition
authority boundary
uncertainty
outer-loop obligation

This prevents ordinary concession logic from treating everything as exchangeable.

Adjudication

The strongest negotiation application is pre-negotiation field clarification and post-agreement witness.

---

13. Social-choice theory

Social-choice theory studies methods for aggregating individual rankings, judgments, or welfare into collective outcomes.

Arrow's impossibility theorem establishes that no social-welfare ordering based on unrestricted ordinal preferences can satisfy a set of plausible conditions simultaneously without dictatorship.

Other major results and methods address:

• majority cycles;
• strategic voting;
• agenda control;
• interpersonal comparisons;
• rights;
• capabilities;
• judgment aggregation.

Strong overlap

• several centers;
• collective choice;
• aggregation rule;
• constitution;
• fairness conditions;
• cycles;
• authority.

Candidate contribution

Polytelometry does not solve Arrow's theorem.

It can make visible:

• which fields should not enter ordinary aggregation;
• which centers have protected standing;
• which alternatives were excluded;
• which rule was selected;
• what unresolved remainder remains after the vote;
• where a fork may be preferable to one collective ordering.

Adjudication

Social-choice limits are a hard boundary against claims that the system can discover one neutral collective telos.

---

14. Deliberative democracy

Deliberative-democracy traditions emphasize reason-giving, reflection, participation, mutual justification, and preference transformation rather than simple aggregation of fixed preferences.

Actual deliberative systems vary widely in:

• selection;
• facilitation;
• power;
• scale;
• information;
• decision authority;
• public connection.

Strong overlap

• preference revision;
• reasons;
• shared field;
• learning;
• legitimacy;
• public standing;
• plural output.

Candidate difference

Polytelometry provides a traceable field and route record.

It emphasizes:

• source versus summary;
• missing standing;
• protected dissent;
• consent and authority;
• route lifecycle;
• model mediation.

Adjudication

Polytelometric navigation is compatible with deliberation but cannot manufacture deliberative legitimacy through software alone.

Selection, power, public authority, and implementation remain institutional questions.

---

15. Polis and scalable deliberation

Polis uses opinion clustering and visualization to support large-scale deliberative participation. Research on LLM augmentation identifies opportunities for facilitation and summarization along with risks from context limitations and system design.

More recent large-scale evaluation work reports persistent problems with minority underrepresentation, order effects, and weak alignment between LLM judges and human judgments in deliberation summarization.

Strong overlap

• large-scale free-form input;
• clustering;
• consensus discovery;
• summary;
• public meaning-making;
• minority representation.

Candidate addition

A polytelometric deliberation output should preserve separate fields:

shared
compatible
conflicting
protected minority
unresolved
missing standing
route proposals

Adjudication

Consensus should never be the only target metric for AI-assisted deliberation.

Representation quality must be judged partly from the perspective of the represented participants.

---

16. Human–AI deliberation

Recent Human–AI Deliberation research proposes structured interaction where humans and AI discuss dimension-level disagreements rather than merely accepting or rejecting a model recommendation.

An exploratory study in graduate-admissions decision-making reported improvements in appropriate reliance and task performance relative to conventional explainable-AI assistance.

Strong overlap

• dimension-level disagreement;
• conversational clarification;
• model and human opinion;
• reflection;
• decision update;
• appropriate reliance.

Candidate difference

Polytelometry requires:

• source and inference separation;
• participant standing;
• consent and authority;
• protected conditions;
• cost bearers;
• route alternatives;
• stop conditions;
• model-provider disclosure.

Adjudication

Human–AI deliberation is the strongest direct implementation neighbor for F.9.

The next technical paper should engage it explicitly.

---

17. Mixed-initiative systems

Mixed-initiative interaction allocates initiative dynamically among people and machines.

A machine may:

• ask;
• suggest;
• complete;
• warn;
• retrieve;
• plan;
• execute.

The central design problem is not only capability.

It is control, timing, interruptibility, and appropriate division of labor.

Telic relevance

A navigation assistant may generate routes while a human controls admission and protected status.

A model may identify a contradiction while an authorized group selects the decision rule.

Initiative should be bounded by:

• role;
• stakes;
• confidence;
• reversibility;
• consent;
• authority;
• ability to contest.

Adjudication

Mixed initiative should be represented as explicit delegated authority, not invisible interface behavior.

---

18. Appropriate reliance

Human–AI decision research distinguishes:

• overreliance;
• underreliance;
• appropriate reliance.

Explanations do not automatically create appropriate reliance. Interface, confidence, task, user skill, framing, and model behavior all matter.

Polytelometric concern

Appropriate reliance must include:

• reliance on what;
• for which center;
• under which authority;
• with what ability to correct;
• at which lifecycle stage.

A user may appropriately rely on the model to organize arguments and inappropriately rely on it to set moral weights.

Adjudication

Role-specific reliance is more useful than one global trust measure.

---

19. Representation versus decision

A recurring architecture distinction should be:

REPRESENT
STRUCTURE
GENERATE
COMPARE
RECOMMEND
AUTHORIZE
EXECUTE
WITNESS

Different agents may hold different roles.

A model may be permitted to:

• represent candidate fields;
• structure records;
• generate routes;
• compare modeled consequences.

It may lack authority to:

• authorize;
• execute;
• determine protected standing.

Adjudication

Role separation is a central Phase I requirement.

It should be machine-readable and visible to participants.

20. Representation fairness

A decision-support system may accurately summarize the majority while misrepresenting the field.

Representation fairness includes:

• source recognition;
• minority retention;
• equal opportunity for correction;
• visibility of uncertainty;
• distinction between frequency and protected standing;
• resistance to order effects;
• resistance to articulate-speaker bias;
• treatment of absent or delegated centers.

A summary can be neutral in tone and still be unfair in structure.

For example, it may give equal textual space to views that differ greatly in affected standing or compress a protected minority concern into “some participants disagreed.”

Adjudication

Representation fairness should be evaluated separately from recommendation accuracy.

This is especially important for LLM-mediated deliberation.

---

21. Agenda and option control

Decision outcomes depend on:

• which question is asked;
• which alternatives are available;
• which sequence is used;
• which defaults apply;
• which issues are bundled;
• when the process ends.

Social-choice and negotiation research already demonstrate the power of agenda and option structure.

Polytelometric addition

The navigation record should preserve:

• who framed the question;
• who generated the alternatives;
• which options were rejected before deliberation;
• what default governs no decision;
• whether participants could propose a fork, pause, or release.

Adjudication

Route generation and agenda governance are as important as route ranking.

A fair decision rule over an unfair option set does not create legitimate navigation.

---

22. Refusal, abstention, and no-decision states

Many formal decision systems assume that one option must be selected.

Legitimate outcomes may include:

• abstain;
• defer;
• refuse;
• preserve status temporarily;
• narrow scope;
• seek another authority;
• fork;
• dissolve;
• declare no adequate route.

Arrow-style formal models often assume a social choice must be made from a feasible set. Operational systems similarly reward completion.

Polytelometric navigation explicitly preserves no-decision states.

Adjudication

A system's ability to stop is part of its decision competence.

The default state and cost of delay must remain visible.

No decision can itself impose cost.

---

23. Route portfolios

One route may not serve every center or scenario.

A route portfolio may combine:

• several actions;
• staged alternatives;
• different local routes;
• contingency plans;
• reversible experiments;
• protected exceptions;
• parallel forks.

Portfolio methods are common in finance, project selection, policy, and robust planning.

Candidate contribution

Polytelometry may recommend plural routes when one shared route would require unnecessary domination.

Examples:

• different service channels;
• local policy variation;
• opt-in governance;
• interoperable forks;
• staged deployment;
• exceptions preserving protected standing.

Boundary

Plural routes can increase:

• cost;
• complexity;
• inequality;
• fragmentation;
• interoperability burden.

Adjudication

Plural-route design should be compared with forced uniformity and with ungoverned fragmentation.

---

24. Adaptive routes

An adaptive route changes according to observed conditions.

It may include:

initial action
signpost
threshold
review
branch
fallback
exit

Adaptive pathways are common in robust planning.

Telic requirements

Each branch should preserve:

• authority;
• consent;
• represented standing;
• trigger provenance;
• correction;
• versioned witness.

An initial consent to a program should not be treated as unlimited consent to every future branch.

Adjudication

Adaptive consent must match adaptive planning.

Route changes can require renewed authorization even when they were technically anticipated.

---

25. Institutional implementation

A decision record is not an implemented route.

Implementation requires:

• capability;
• authority;
• resources;
• translation;
• timing;
• accountability;
• monitoring;
• repair;
• succession.

A route may be selected through an excellent deliberative process and fail through institutional incompatibility.

F.7's substrate and compatibility analysis therefore follows F.9.

Adjudication

Polytelometric navigation should link each route to:

• implementation substrate;
• responsible loop;
• dependency;
• capacity;
• dissolution or rollback path.

The chosen route should not become an orphaned recommendation.

---

26. Decision witness and learning

The navigation system should compare:

projected consequence
actual consequence
represented standing
actual cost bearer
expected consent
actual experience
declared telos
operative result

Decision quality cannot be assessed only at selection time.

A route may look balanced and later reveal:

• hidden burden;
• model error;
• false assumption;
• preference change;
• institutional capture;
• unexpected benefit.

The witness should support:

• correction;
• restitution;
• model update;
• rule revision;
• route release;
• preserved dissent that later proved material.

Adjudication

A decision method should be evaluated partly by how well it learns from its own field errors.

---

27. Comparison matrix

| Method | Primary strength | Typical limitation addressed by polytelometry | |---|---|---| | Multiobjective optimization | Efficient solution sets across objectives | Objective admission, standing, protected fields | | Pareto decision support | Exploration of nondominated alternatives | Authority, consent, and omitted objectives | | MCDA | Structured criteria, values, weighting, sensitivity | Source status, boundaries, lifecycle | | Outranking | Partial order, veto, incomparability | Standing basis and review of veto | | Value-focused thinking | Values-first route generation | Several centers, consent, outer-loop effects | | Problem structuring | Contested framing and wicked problems | Versioned telic records and authority gates | | Robust decision-making | Vulnerability across plausible futures | Robust for whom, protected conditions | | Scenario planning | Strategic learning across futures | Scenario provenance and absent standing | | Argumentation | Claims, attacks, support, acceptability | Non-argumentative standing and action authority | | Negotiation support | Offers, concessions, agreement | Protected conditions and coercive dependence | | Social choice | Aggregation rules and formal limits | Nonaggregable fields, unresolved remainder | | Deliberative democracy | Preference formation and public justification | Traceable field, route, and model mediation | | Polis | Scalable clustering and consensus discovery | Minority and protected-field preservation | | Human–AI deliberation | Structured disagreement and appropriate reliance | Source, consent, cost bearer, stop | | Mixed initiative | Dynamic division of labor | Explicit delegated authority and lifecycle |

---

28. Terminology adjudication

Polytely

Decision: use as established prior term.

Boundary: presence or interaction of multiple goals.

Polytelometry

Decision: retain provisionally.

Boundary: disciplined representation and governance; not universal quantitative measurement.

Polytelometric navigation

Decision: retain.

Boundary: pre-optimization and lifecycle layer, not a universal decision rule.

Field class

Decision: retain.

Candidate classes:

shared
compatible
conditional
conflicting
protected
unresolved
missing-standing
released

Route

Decision: retain.

Boundary: sequence, gates, review, and exit—not endpoint alone.

Protected field

Decision: retain cautiously.

Boundary: nontradeable under current authority; not eternally absolute.

Missing-standing field

Decision: retain.

Cost-bearer map

Decision: adopt.

Weight provenance

Decision: adopt.

Route portfolio

Decision: retain provisionally.

Adaptive consent

Decision: adopt provisionally.

No-decision state

Decision: adopt as a first-class outcome.

Representation fairness

Decision: adopt as a research dimension.

Decision witness

Decision: retain.

Constitutional Pareto review

Decision: avoid as a public coinage for now.

Prefer:

Pareto analysis after standing and protected-condition review.

---

29. Claim adjudication

May be stated strongly

• Many decisions involve several objectives and centers.
• Pareto efficiency is relative to the represented objectives and feasible alternatives.
• Scalarization embeds value judgments.
• MCDA supports structured comparison but does not remove judgment.
• Some multiobjective sequential problems should not be reduced to one scalar objective.
• Robust decision-making supports choices under deep uncertainty.
• Argumentation models reasons and attacks but not all forms of standing.
• Social-choice theory establishes formal limits on preference aggregation.
• Deliberation may change preferences and problem framing.
• LLM summaries can underrepresent minority positions.
• Human–AI deliberation can support reflection and appropriate reliance in some studied tasks.
• Decision outputs depend upon agenda, option, and default structure.

May be stated as a proposed synthesis

• Telic admission should precede objective formalization.
• Field classes improve plural-end representation.
• Cost-bearer maps expose externalization.
• Protected fields should remain outside ordinary scalarization.
• Polytelometric records can govern MCDA, MOO, deliberation, and negotiation.
• No-decision states should be first-class outcomes.
• Adaptive routes require adaptive consent.
• Representation fairness should be evaluated separately from outcome accuracy.
• Decision witness should link routes to later consequences.

Must remain hypotheses

• Polytelometric pre-navigation improves decision legitimacy.
• Field classification can be performed reliably.
• Route diversity improves outcomes enough to justify additional burden.
• Cost-bearer maps change externalizing decisions.
• Protected-field status can be governed without strategic abuse.
• LLM assistants can preserve minority and unresolved fields at scale.
• The navigation record scales to public deliberation.
• Explicit stop competence reduces harm without causing excessive paralysis.

Should not be claimed

• Polytelometry solves Arrow's theorem.
• A Pareto-efficient route is legitimate.
• One objective set captures the whole field.
• Weights objectively measure moral importance.
• Consensus proves shared purpose.
• Deliberation eliminates power.
• A model can neutrally summarize every voice.
• Protected fields always override all other standing.
• More routes always improve a decision.
• Robustness eliminates uncertainty.
• Human authorization automatically resolves absent standing.

---

30. Required F.9 boundaries

F.9 appropriately includes:

• polytely versus polytelometry;
• navigation versus optimization;
• source and telic-item types;
• field classes;
• route generation before ranking;
• tradeoff authority;
• cost bearers;
• reversibility;
• Pareto and scalarization boundaries;
• robustness and scenarios;
• argument, deliberation, social choice, and negotiation;
• consent and authority gates;
• witness;
• stop and escalation;
• model non-sovereignty;
• lifecycle and candidate record;
• falsification and ethical boundaries.

Before Phase I use, the record requires:

• integration with the Projection Record;
• integration with the Capacity Profile;
• integration with Temporal Standing;
• integration with Semantic Trail and PROV;
• usability and burden testing;
• formal representation of protected and unresolved fields;
• governance of weights and vetoes;
• benchmark design for representation fairness.

---

31. Empirical differentiation agenda

31.1 MCDA extension study

Add standing, evidence, authority, consent, and protected status to an MCDA workflow.

Measure whether route selection changes.

31.2 Pareto admission study

Generate a Pareto front before and after affected-center and protected-condition review.

Compare the resulting frontier and selected solutions.

31.3 Deliberation summarization benchmark

Evaluate summaries on:

• majority representation;
• minority recognition;
• protected concern;
• source recognition;
• uncertainty;
• route diversity;
• participant correction.

31.4 Social-choice boundary study

Test when participants prefer:

• vote;
• consensus;
• supermajority;
• veto;
• fork;
• no decision.

Record the constitutional assumptions behind the selected rule.

31.5 Human–AI role study

Compare model roles:

recommend
explain
deliberate
navigate
execute

Measure reliance, authority confusion, correction, and route diversity.

31.6 Adaptive-consent study

Test whether users understand and accept branching plans when consent is renewed at signposts versus granted once at entry.

31.7 Decision-witness study

Measure whether preserved rationale and dissent improve later review, repair, and learning.

---

32. Candidate Phase I architecture implications

G.10 supports a modular design.

32.1 Admission layer

Consumes:

• centers;
• projections;
• evidence status;
• protected conditions;
• temporal status.

32.2 Field classifier

Produces:

• shared;
• compatible;
• conditional;
• conflicting;
• protected;
• unresolved;
• missing-standing;
• released.

32.3 Route generator

Human and machine route proposals remain attributed.

32.4 Analysis adapters

Adapters may invoke:

• Pareto analysis;
• MCDA;
• outranking;
• robust scenario analysis;
• argumentation;
• voting;
• negotiation.

32.5 Governance gates

• standing;
• consent;
• authority;
• capacity;
• privacy;
• stop.

32.6 Decision rule declaration

The selected aggregation or choice rule must be explicit.

32.7 Witness layer

Stores:

• source;
• transformation;
• route;
• rule;
• dissent;
• consequence;
• review.

32.8 Lifecycle controller

Supports:

• expiry;
• review;
• revision;
• fork;
• release;
• dissolution.

The architecture should not contain a hidden universal optimizer.

---

33. Bottom line

F.9 should proceed.

It should proceed as:

• a governance and problem-structuring layer;
• a bridge among MOO, MCDA, robust planning, argumentation, negotiation, social choice, and deliberation;
• a route-lifecycle method;
• a source of measurable representation and authority requirements;
• a foundation for non-sovereign human–AI navigation.

It should not proceed as:

• a new universal optimization algorithm;
• a solution to social-choice impossibility;
• a promise of neutral aggregation;
• a requirement to quantify every value;
• a substitute for public institutions;
• a system that treats every disagreement as negotiable;
• a method for models to assign moral weights.

The strongest defensible formulation is:

Polytelometric navigation extends established decision-support methods by governing the admission, status, authority, and lifecycle of the ends those methods operate upon, preserving protected and unresolved fields that should not be silently reduced to weights, votes, or model summaries.

---

Primary references

Arrow, Kenneth J. Social Choice and Individual Values. Wiley, 1951; revised edition, Yale University Press, 1963.

Belton, Valerie, and Theodor J. Stewart. Multiple Criteria Decision Analysis: An Integrated Approach. Springer, 2002.

Dung, Phan Minh. “On the Acceptability of Arguments and Its Fundamental Role in Nonmonotonic Reasoning, Logic Programming and n-Person Games.” Artificial Intelligence 77, no. 2, 1995.

Keeney, Ralph L. Value-Focused Thinking: A Path to Creative Decisionmaking. Harvard University Press, 1992.

Lempert, Robert J., Steven W. Popper, and Steven C. Bankes. Shaping the Next One Hundred Years: New Methods for Quantitative, Long-Term Policy Analysis. RAND Corporation, 2003.

Ma, Shuai, et al. “Towards Human-AI Deliberation: Design and Evaluation of LLM-Empowered Deliberative AI for AI-Assisted Decision-Making.” arXiv:2403.16812, 2024.

Osika, Zuzanna, et al. “What Lies beyond the Pareto Front? A Survey on Decision-Support Methods for Multi-Objective Optimization.” arXiv:2311.11288, 2023.

Rittel, Horst W. J., and Melvin M. Webber. “Dilemmas in a General Theory of Planning.” Policy Sciences 4, 1973.

Roijers, Diederik M., Peter Vamplew, Shimon Whiteson, and Richard Dazeley. “A Survey of Multi-Objective Sequential Decision-Making.” Journal of Artificial Intelligence Research 48, 2013.

Sen, Amartya. Collective Choice and Social Welfare. Holden-Day, 1970.

Small, Christopher T., et al. “Opportunities and Risks of LLMs for Scalable Deliberation with Polis.” arXiv:2306.11932, 2023.

Zhu, Shenzhe, et al. “Can AI Truly Represent Your Voice in Deliberations? A Comprehensive Study of Large-Scale Opinion Aggregation with LLMs.” arXiv:2510.05154, 2025.