Reasoning Playbook · Glossary

Marketing Reasoning Glossary

How reasoning shows up in marketing copy — cause-effect claims, analogies, comparisons, decision logic, objection-handling, and narrative. A subset of the full Reasoning Playbook, curated to the moves that actually fire on landing pages, product copy, and brand voice.

33 modules

B) Ampliative reasoning (conclusions go beyond the premises)

ampliative:inductive

Inductive reasoning (generalization)

#
Infer general patterns from observations ("observed many A are B -> probably A are B").
Outputs
General rules, trends, predictors.
How it differs
Not truth-preserving; new data can overturn it.
Best for
Learning from experience, early-stage pattern discovery, forming priors.
Failure mode
Overgeneralizing from small/biased samples.
Pairs withmetalevel:adversarial to stress-test the rule, ampliative:analogical to check if the pattern holds in other domains.
ampliative:statistical

Statistical reasoning (frequentist style)

#
Inference about populations from samples via estimators, confidence intervals, tests, error rates.
Outputs
Effect estimates + uncertainty statements tied to sampling procedures.
How it differs
Typically avoids "probability of hypotheses"; emphasizes long-run properties of procedures.
Best for
Experiments, A/B tests, QA, inference under repeated-sampling assumptions.
Failure mode
P-value worship; confusing "no evidence" with "evidence of no effect."
Pairs withuncertainty:bayesian for complementary inference, domain:experimental for design.
ampliative:abductive

Abductive reasoning (inference to the best explanation)

#
From observations, propose a hypothesis that would best explain them.
Outputs
Candidate explanations/models; "best current story."
How it differs
Unlike induction (generalizing frequencies), abduction introduces hidden mechanisms/causes; unlike deduction, it's not guaranteed.
Best for
Hypothesis generation, incident triage, diagnosis, scientific discovery.
Failure mode
"Story bias" (choosing the most appealing explanation, not the most supported).
Pairs withdomain:experimental to design the test, research:evidence-table to ground hypotheses in sources.
ampliative:analogical

Analogical reasoning (structure mapping)

#
Transfer relational structure from a known domain/case to a new one (often deeper than surface similarity).
Outputs
Candidate inferences; adapted solutions; conceptual models/metaphors.
How it differs
Often particular -> particular transfer; frequently seeds abduction ("maybe it works like...").
Best for
Innovation, design, teaching, cross-domain problem solving.
Failure mode
False analogies (shared surface traits, different causal structure).
Pairs withampliative:inductive to check if the transferred solution generalizes. Audience Playbook: audience:jtbd to test if the analogy resonates with readers; audience:mental-models to verify the source domain is one the audience already inhabits.
ampliative:case-based

Case-based reasoning (exemplar retrieval + adaptation)

#
Retrieve similar past cases and adapt their solutions.
Outputs
Proposed solution justified by precedent; playbook actions.
How it differs
More operational than analogy: emphasizes retrieval metrics + adaptation operators + case libraries.
Best for
Law (precedent), customer support, clinical decision support, ops playbooks.
Failure mode
Cargo-culting: applying precedent without checking context changes.
Pairs withampliative:analogical for deeper structural mapping, revision:defeasible for handling exceptions.
ampliative:explanation-based

Explanation-based learning / reasoning

#
Use an explanation of why a solution works to generalize a reusable rule/plan.
Outputs
Generalized strategies with an explanatory justification.
How it differs
It generalizes like induction but is guided/validated by deductive explanation.
Best for
Turning expert solutions into SOPs; reducing overfitting to anecdotes.
Failure mode
Explanations that are internally elegant but empirically wrong.
Pairs withampliative:inductive for pattern validation, causal:mechanistic for deeper explanation.
ampliative:simplicity

Simplicity / compression reasoning (Occam, MDL)

#
Prefer hypotheses that explain data with fewer assumptions / shorter descriptions, balancing fit vs complexity.
Outputs
Bias toward simpler models; complexity penalties; regularization choices.
How it differs
It's a selection principle across hypotheses; often paired with abduction and statistics.
Best for
Model selection, avoiding overfitting, choosing parsimonious policies.
Failure mode
Oversimplifying when the world is genuinely complex/nonlinear.
Pairs withampliative:abductive for generating candidates, uncertainty:bayesian for formal model comparison.
ampliative:reference-class

Reference-class / "outside view" reasoning

#
Predict by comparing to a base rate distribution of similar past projects/cases ("what usually happens?").
Outputs
Base-rate forecasts; adjustment factors.
How it differs
It's an inductive method designed to counter planning fallacy and inside-view optimism.
Best for
Project timelines, budgets, risk forecasting, portfolio-level planning.
Failure mode
Choosing the wrong reference class (too broad or too narrow).
Pairs withmetalevel:calibration for confidence checking, uncertainty:bayesian for formal updating.

D) Reasoning under vagueness and borderline concepts (graded predicates)

vagueness:prototype

Prototype / similarity-based category reasoning

#
Categorize by similarity to prototypes/exemplars rather than strict necessary-and-sufficient definitions.
Outputs
Graded category judgments; typicality effects.
How it differs
Natural for human categories; complements fuzzy/rough by focusing on similarity geometry.
Best for
UX taxonomies, product categorization, human-facing labeling.
Failure mode
Hidden bias in prototypes; category drift over time.
Pairs withvagueness:fuzzy for graded membership, ampliative:analogical for cross-domain categorization.

E) Reasoning with inconsistency, defaults, and changing information

revision:defeasible

Defeasible reasoning (tentative conclusions + defeat relations)

#
Conclusions can be defeated by counterevidence or stronger rules; tracks priorities/strength.
Outputs
Warranted conclusions given competing reasons.
How it differs
More explicit about conflict resolution than plain defaults.
Best for
Compliance/policy, medical guidelines, conflicting requirements.
Failure mode
Priority schemes that encode politics rather than relevance.
Pairs withdialectical:argumentation for structured pro/con, revision:belief-revision for principled updating.
dialectical:argumentation

Argumentation theory (structured pro/con evaluation)

#
Build arguments and counterarguments; compute which claims stand given attack/defense relations.
Outputs
Accepted/warranted claims; rationale maps.
How it differs
Not just "derive consequences" but "evaluate competing reasons."
Best for
Governance, policy disputes, legal-style reasoning, stakeholder conflicts.
Failure mode
Mistaking "won the debate" for "is true" (argument strength vs reality).
Pairs withrevision:defeasible for priority handling. Rhetoric Playbook: appeal:logos for the logical structure, appeal:ethos for the credibility framing of an argumentation case.

F) Causal, counterfactual, explanatory, and dynamic reasoning

causal:inference

Causal inference (interventions vs observations)

#
Identify causal relations and predict effects of interventions (distinguish P(Y|X) vs P(Y|do(X))).
Outputs
Causal effect estimates; intervention predictions; adjustment sets.
How it differs
Correlation alone can't resolve confounding or direction; causal reasoning encodes structure assumptions.
Best for
Product impact, policy evaluation, root-cause analysis that must guide action.
Failure mode
Hidden confounders; unjustified causal assumptions.
Pairs withdomain:experimental to design the test, causal:counterfactual to trace intervention consequences.
causal:counterfactual

Counterfactual reasoning ("what would have happened if...")

#
Evaluate alternate histories given a causal model.
Outputs
Counterfactual outcomes; blame/credit analyses; individualized explanations.
How it differs
Needs causal structure beyond pure statistics.
Best for
Postmortems, accountability, scenario evaluation, personalized decision support.
Failure mode
Confident counterfactuals from weak models.
Pairs withcausal:inference for the underlying model, causal:mechanistic for understanding the propagation.
causal:mechanistic

Mechanistic reasoning (how it works internally)

#
Explain/predict by identifying parts and interactions.
Outputs
Mechanistic explanations; levers; failure modes.
How it differs
Stronger than correlation: gives actionable intervention points and generalizes when mechanisms hold.
Best for
Engineering, debugging, safety analysis, biology/medicine.
Failure mode
"Just-so mechanisms" that sound plausible but aren't validated.
Pairs withcausal:inference for effect estimation, causal:systems-thinking for feedback dynamics.
causal:diagnostic

Diagnostic reasoning (effects -> causes under constraints)

#
Infer hidden faults/causes from symptoms using a fault/causal model plus uncertainty handling.
Outputs
Ranked causes; next-best tests; triage plans.
How it differs
Often abduction + Bayesian/likelihood updates, constrained by explicit fault models.
Best for
Incident response, troubleshooting, quality triage.
Failure mode
Premature closure (locking onto one cause too early).
Pairs withcausal:mechanistic for understanding failure modes, practical:value-of-information for test prioritization.

G) Practical reasoning (choosing actions under constraints)

practical:means-end

Means-end / instrumental reasoning

#
From goals, derive actions/subgoals necessary or helpful to achieve them ("to get X, do Y").
Outputs
Action rationales; subgoals; dependency chains.
How it differs
About doing, not merely believing; feeds planning and decision theory.
Best for
Strategy decomposition, OKRs, operational planning.
Failure mode
Local means become ends ("process is the goal").
Pairs withformal:deductive for forward verification of the plan. Audience Playbook: audience:barriers to identify obstacles to reader engagement with the plan.
practical:decision

Decision-theoretic reasoning (utilities + uncertainty)

#
Combine beliefs with preferences/utilities to choose actions (e.g., expected utility).
Outputs
Option rankings; policies; explicit tradeoffs.
How it differs
Bayesian reasoning updates beliefs; decision theory adds values and consequences.
Best for
Portfolio choices, risk decisions, prioritization, pricing.
Failure mode
Utility mismatch (what you optimize isn't what you truly value).
Pairs withuncertainty:bayesian for probability estimates, practical:robust for worst-case analysis.
practical:multi-criteria

Multi-criteria decision analysis (MCDA) / Pareto reasoning

#
Decide with multiple objectives (cost, speed, safety, equity), often using weights, outranking, or Pareto frontiers.
Outputs
Tradeoff surfaces; Pareto-efficient sets; transparent scoring models.
How it differs
Makes tradeoffs explicit instead of collapsing them implicitly into one objective.
Best for
Strategy, procurement, roadmap planning, governance.
Failure mode
Arbitrary weights hiding politics; false precision.
Pairs withstrategic:negotiation for stakeholder alignment, practical:decision for expected-value analysis.
practical:satisficing

Satisficing (bounded rationality with stopping rules)

#
Seek a solution that is "good enough" given time/compute/info limits rather than globally optimal.
Outputs
Thresholds; stopping rules; acceptable solutions.
How it differs
Not "lazy optimization"; it's rational under constraints.
Best for
Real-time ops, fast-moving environments, early product strategy.
Failure mode
Thresholds too low leads to chronic mediocrity; too high leads to disguised optimization.
Pairs withpractical:optimization when more time is available, metalevel:meta-reasoning for deciding the right effort level.
practical:heuristic

Heuristic reasoning (fast rules of thumb)

#
Use simple rules that often work; fast but biased.
Outputs
Quick decisions/inferences; prioritization shortcuts.
How it differs
Less principled but cheaper; should be paired with checks/calibration.
Best for
Triage, first drafts, guiding search.
Failure mode
Heuristics become doctrine.
Pairs withmetalevel:debiasing for checking blind spots, metalevel:calibration for confidence assessment.

H) Strategic and social reasoning (other agents matter)

strategic:theory-of-mind

Theory-of-mind / mental-state reasoning

#
Infer beliefs, intentions, knowledge states of others (nested beliefs).
Outputs
Behavior predictions; communication strategies; coordination plans.
How it differs
Focuses on beliefs-about-beliefs; often essential for collaboration.
Best for
Leadership, UX, teamwork, threat modeling.
Failure mode
Mind-reading with overconfidence; projecting your incentives onto others.
Pairs withstrategic:game for strategic interaction. Audience Playbook: audience:empathy-map for reader modeling; audience:mental-models for the cognitive frame the reader brings.
strategic:negotiation

Negotiation and coalition reasoning

#
Reason about acceptable agreements and coalition formation under constraints and asymmetric information.
Outputs
Offers, concessions, coalition structures; Pareto improvements.
How it differs
More process- and constraint-oriented than abstract equilibrium analysis; mixes game theory with norms/rhetoric.
Best for
Partnerships, sales, cross-team alignment.
Failure mode
Winning the negotiation but losing the relationship/long-term incentives.
Pairs withstrategic:theory-of-mind for predicting reactions, practical:multi-criteria for tradeoff analysis.

I) Dialectical, rhetorical, and interpretive reasoning (reasoning as a human practice)

dialectical:hermeneutic

Hermeneutic / interpretive reasoning (meaning under ambiguity)

#
Infer meaning and intent from language, documents, norms, artifacts using context and interpretive canons.
Outputs
Interpretations; reconciled meanings; clarified definitions.
How it differs
Emphasizes context and ambiguity management, not only formal entailment.
Best for
Contracts, policy docs, requirements, qualitative feedback synthesis.
Failure mode
Over-interpreting; reading intent that isn't there.
Pairs withvagueness:partial-logic for handling indeterminacy, dialectical:argumentation for competing interpretations. Rhetoric Playbook: dialectical:charity for the principle of charitable interpretation as a writing move.
dialectical:narrative

Narrative reasoning / causal storytelling

#
Build coherent time-ordered explanations connecting events, motives, causes into a story supporting prediction and action.
Outputs
Postmortems, strategy narratives, scenario stories.
How it differs
Integrates causal/abductive/rhetorical constraints; risk is over-coherence ("too neat").
Best for
Incident reports, executive communication, explaining complex causal chains.
Failure mode
Narrative closure crowding out alternative hypotheses.
Pairs withcausal:counterfactual for alternative histories. Rhetoric Playbook: frame:narrative for the writing-frame that this reasoning produces; frame:paradox when the causal story turns on a tension.
dialectical:sensemaking

Sensemaking / frame-building reasoning

#
Decide "what kind of situation is this?" -- build frames that organize signals, priorities, and actions under ambiguity.
Outputs
Situation frames; working hypotheses; shared mental models.
How it differs
Precedes many other modes: it selects what counts as relevant evidence and what questions to ask.
Best for
Crisis leadership, early-stage strategy, ambiguous competitive landscapes.
Failure mode
Locking onto the wrong frame and then reasoning flawlessly inside it.
Pairs withmetalevel:meta-reasoning for choosing how to reason, ampliative:abductive for hypothesis generation. Rhetoric Playbook: dialectical:reframe to move a debate from one frame to another in writing; frame:identity when the situation is best understood through who-the-reader-is.

J) Modal, temporal, spatial, and normative reasoning (structured possibility, time, space, and "ought")

K) Domain-specific reasoning styles (practice changes the "rules")

domain:moral

Moral / ethical reasoning

#
Reason about right/wrong and value tradeoffs (consequentialist, deontological, virtue, contractualist, care ethics, etc.).
Outputs
Value constraints; ethical justifications; tradeoff statements.
How it differs
Normative: cannot be reduced to facts alone, though must be informed by them.
Best for
AI governance, product harms, trust & safety, people policy.
Failure mode
Values laundering ("it's 'ethical' because it helps our goal") without principled constraints.
Pairs withmodal:deontic for norm analysis, metalevel:reflective-equilibrium for coherence checking.

L) Meta-level and reflective modes (reasoning about reasoning)

metalevel:calibration

Calibration and epistemic humility (second-order uncertainty)

#
Track how reliable your beliefs are (forecast scoring, error bars, backtesting).
Outputs
Calibrated confidence; forecast accuracy metrics; improved priors.
How it differs
First-order uncertainty is "what is true?"; calibration is "how good am I at knowing?"
Best for
Forecasting culture, risk reviews, decision reviews.
Failure mode
Confusing confidence with competence; never measuring accuracy.
Pairs withquality:assumption-audit for deeper challenge, research:uncertainty-question for follow-up.
metalevel:adversarial

Adversarial / red-team reasoning

#
Assume the role of an attacker/critic: try to break arguments, systems, incentives, and assumptions.
Outputs
Failure modes, exploits, counterexamples, "what could go wrong" maps.
How it differs
It's intentionally antagonistic to your current plan; pairs with robust reasoning and assurance cases.
Best for
Security, safety, governance, strategy stress-testing.
Failure mode
Cynicism theater (finding clever attacks without prioritizing real risk).
Pairs withformal:deductive to repair the argument. Rhetoric Playbook: dialectical:steelman to construct the strongest opposing argument; argument:concession-refute to integrate it into prose.

Q) Quality modules

quality:assumption-audit

Assumption Audit

#
Surface and challenge every assumption in an analysis.
Outputs
Assumption inventory with impact ranking and verification priorities.
How it differs
Explicit assumption extraction followed by impact assessment.
Best for
Final check on any analysis before committing to its conclusions.
Failure mode
Listing obvious assumptions while missing the dangerous hidden ones.
Pairs withmetalevel:adversarial for assumption attack, metalevel:calibration for confidence calibration.
quality:calibration

Calibration Check

#
Assess whether confidence levels in the analysis are warranted.
Outputs
Calibration assessment with specific adjustments for each claim.
How it differs
Match claim strength to evidence strength -- are you overclaiming or underclaiming?
Best for
Posts that make quantitative or strength-of-evidence claims.
Failure mode
Calibrating language without calibrating substance.
Pairs withsynthesis:claims for integration, metalevel:calibration for overall confidence.
No modules match that filter.