Public Speeches

Public Speeches

Summary

Public speeches convey grievances, demands, or values to a live or mediated audience. In Planning Mode, this skill produces a preparation package: audience model, speaker rationale, research notes, message brief, outline choice, risk review, amplification notes, rehearsal criteria, and debrief prompts. In Application Mode, it carries that preparation through to a speakable final speech package. A speech can support campaign narrative infrastructure, but method-level effectiveness evidence remains thin; evaluate it as one communication tactic within a broader context, not as a standalone lever.

When to Consult This Skill

Consult when any of the following conditions apply:

• a user wants help preparing a public address, rally speech, kickoff speech, statement, or spoken appeal
• a user asks what the speech should include, what to research, or how to make the speech legible to a specific audience
• a user needs a structured speech brief before drafting or rehearsing
• a user or game runtime asks to draft a complete speech or apply the tactic end-to-end
• a planner needs to assess speaker legitimacy, venue fit, documentation, media handling, or endpoint
• a user asks how to reduce legal, narrative, reputational, or crowd-control risk around public oratory

Core Mechanism

The tactic operates through declarative communication. A speaker states a position in public, and the statement becomes citable, recordable evidence that a constituency holds that position. Connect the mechanism to design choices:

• Audience movement. Name who needs to understand, trust, care, or act differently.
• Amplification. Shape one accurate line that witnesses, press, and social channels can repeat without distorting the point.
• Proof-of-presence. Align speaker, convenor, audience, venue, and documentation so the claim has visible standing.

The speech's force depends less on rhetorical style than on fit: speaker attribution, claimed constituency, audience, evidence, timing, and channel.

Typical Roles and Participants

Identify which roles exist, which are missing, and which are intentionally compressed:

• Speaker. Voice attached to the claim; legitimacy depends on relationship to the constituency and issue.
• Convenor / host. Organization, venue, or informal group lending infrastructure and public standing.
• Primary audience. People in the room or live channel whose perception or commitment matters first.
• Secondary audience. Press, social networks, institutional observers, opponents, and later viewers.
• Documentarian. Person or team capturing usable stills, full video, short clips, and accurate records.
• Media lead. Person handling press context, quote discipline, inbound questions, and post-event materials.
• Legal / safety observer. Relevant when venue, policing, counter-mobilization, or speech restrictions create risk.
• Endpoint owner. Person responsible for what listeners do after the speech ends.

Roles are descriptive, not mandatory. Small events may compress several roles into one person.

Enabling Conditions

Speeches tend to be more useful when several conditions are present:

• the user can state the desired listener action in one sentence
• the speaker's attribution matches the constituency or experience being represented
• the audience is segmented into primary, secondary, and institutional listeners
• the core claim can be supported by verifiable facts, lived experience, or named public commitments
• venue, timing, and channel fit the story the speech is meant to carry
• documentation and amplification can preserve the speech's most important line accurately
• legal and reputational boundaries are understood before wording is finalized

Absence of these conditions does not make a speech impossible, but it usually changes what the agent should recommend, research, or flag as uncertain.

Risk Signals and Escalation Triggers

Outcomes tend to be worse when:

• the speaker claims to represent people who have not authorized or recognized that claim
• the core claim depends on unverified allegations, private personal details, defamatory framing, or inflammatory language
• legal thresholds around incitement, defamation, restricted speech, assembly, or venue rules are unclear
• the venue is perceived as captured by one side and delegitimizes the speaker before delivery
• opponents already have a counter-frame that can recontextualize the speech faster than organizers can explain it
• documentation can be confiscated, selectively edited, or stripped of context
• the event has no endpoint, leaving listeners unsure what to do next

Distinguish verified facts, user claims, and assumptions. Avoid inventing citations, escalating rhetoric, or advising evasion of legal or safety constraints.

How to Prepare

Use this skill in the mode that matches the request:

• Planning Mode. Use when the user asks for preparation, research, a brief, risk review, or advice. Output research notes, message brief, outline, risk review, rehearsal feedback criteria, and debrief prompts. Do not draft a full speech unless the user asks.
• Application Mode. Use when the user or game runtime asks to apply the tactic, draft the speech, or produce a final package. Use references/application-protocol.md to follow the phase map from intake and research through full draft, short version, quote-safe line, review, completion ledger, readiness verdict, and final speech package.

Load the reference files as needed:

• Intake and fit. Use references/intake-questionnaire.md to gather goal, audience, speaker standing, issue context, venue, constraints, desired listener action, and likely counter-frame. Ask only for missing information.
• Research. Use references/research-protocol.md when the topic is concrete and public or local research tools are available. Separate verified facts, user claims, and assumptions.
• Message brief. Use references/public-speech-brief-template.md to convert raw context into a one-sentence claim, audience model, values frame, proof points, exclusions, endpoint, and media line.
• Outline. Use references/speech-outline-scaffold.md to choose between public narrative, stump-speech, short statement, or issue-briefing structures.
• Review and learning loop. Use references/risk-review-and-debrief.md before delivery and after the event for legal, narrative, reputational, rehearsal, documentation, and debrief checks.

The agent should preserve source discipline throughout: no invented citations, no inflammatory claims unsupported by evidence, no deceptive attribution, no defamatory wording, no incitement, and no advice to evade lawful venue, safety, or legal constraints.

Sources

Primary (descriptive, taxonomic):

• Beer NVI #104 / Sharp #001: the canonical entry for "Public Speeches" in Sharp's first category (Formal Statements).
• Albert Einstein Institution, 198 Methods of Nonviolent Action.
• Michael Beer / ICNC, Civil Resistance Tactics in the 21st Century.
• Full source index: ../../methods-registry/03-sources-and-references.md

Secondary (preparatory, training-derived, paraphrased):

• Public narrative: Harvard Kennedy School / Marshall Ganz and Harvard DASH worksheet record.
• Campaign and media communication: NDI Campaign Skills Trainer's Guide, The Commons media training, and The Commons campaign communications course.
• Pattern and speaking practice: Beautiful Trouble toolbox guide, Toastmasters speechwriting guidance, CANVAS / ICNC training patterns, Swarthmore case observations, and standard movement media training.

Evidence caveat: ../../civil-resistance-evidence-and-critique.md: method-level effectiveness claims for speeches in isolation outrun the evidence base.

Definition of Done

User-value checks for this skill. Keep boxes unchecked; they are review criteria, not author attestation. Generic repository lint rules live in skills-catalog/README.md.

• Intake is complete, or missing facts are listed for the user to supply or verify.
• Research notes label verified facts, user claims, assumptions, source reliability, and unresolved contradictions.
• The speech brief states the core claim, audience model, speaker rationale, values frame, proof points, exclusions, endpoint, and media line.
• An outline format is chosen for the context, and the outline includes transitions plus a read-aloud test target.
• Legal, narrative, reputational, attribution, privacy, and documentation risks are reviewed before drafting or delivery.
• Rehearsal feedback and post-speech debrief criteria are prepared.
• If Application Mode was requested, the final package includes a full speech draft, short version, quote-safe line, delivery notes, risk review, debrief criteria, completion ledger, and readiness verdict.

Version: 0.8.0 Last Updated: 2026-04-25