DECISION POINTS
Format Selection Tree: Choose structure based on audience and time constraints:
IF legal/judicial context:
├─ IF <10 minutes → Classical 5-part (abbreviated)
└─ IF >10 minutes → Classical 5-part (full)
IF policy/legislative context:
├─ IF formal body → Congressional format
└─ IF advocacy → Policy debate format
IF public/corporate presentation:
├─ IF <15 minutes → TED-style structure
├─ IF hostile audience → Classical (heavy refutatio)
└─ IF neutral/friendly → TED-style or Congressional
IF competitive debate: ├─ IF policy debate → Plan/Inherency/Harms/Solvency └─ IF parliamentary → Classical 5-part
**Argument Ordering Decision:**
- IF hostile audience → Start with shared values, bury weakest point in middle
- IF neutral audience → Primacy/recency (2nd strongest first, strongest last)
- IF friendly audience → Lead with strongest, maintain momentum
**Time Allocation by Format:**
- Classical: Exordium 10% | Narratio 20% | Confirmatio 35% | Refutatio 25% | Peroratio 10%
- TED-style: Story 25% | Problem 25% | Solution 35% | Call-to-Action 15%
- Policy: Plan 15% | Inherency 25% | Harms 30% | Solvency 30%
## FAILURE MODES
**1. Exordium Overload**
- *Symptoms:* Introduction >15% of speech, multiple stories, thesis buried
- *Detection:* If you can't find the thesis in first 60 seconds of a 5-minute speech
- *Fix:* Single hook, quick ethos signal, clear thesis statement. Cut everything else.
**2. Evidence Dumping**
- *Symptoms:* Lists of facts without warrants, no impact statements, audience glazing over
- *Detection:* If confirmatio has >5 discrete evidence pieces without logical connectors
- *Fix:* Group evidence into 2-3 major claims. Use Toulmin: Claim → Warrant → Evidence → Impact
**3. Strawman Refutation**
- *Symptoms:* Addressing weak or irrelevant objections while ignoring the strongest counter
- *Detection:* If refutatio feels easy or opposition sounds unreasonable
- *Fix:* Steel-man the best objection. If you can't make their case sound compelling, you haven't understood it.
**4. Fizzle Finish**
- *Symptoms:* Peroratio restates points without escalation, trails off with qualifications
- *Detection:* Final sentence contains "maybe," "perhaps," or "in conclusion"
- *Fix:* End with imperative verb. Return to opening image. No hedging in final 30 seconds.
**5. Format Mismatch**
- *Symptoms:* Using TED structure for legal argument, or classical structure for policy pitch
- *Detection:* Audience looks confused despite clear content
- *Fix:* Match structure to audience expectations (see decision tree above)
## WORKED EXAMPLES
**Example 1: Legal Closing (200 words)**
[EXORDIUM] "Members of the jury, three weeks ago you promised to follow the evidence wherever it leads. Today, it leads to one conclusion: reasonable doubt."
[NARRATIO] Timeline: incident at 9:47pm → police arrival 10:15pm → witness statements collected 11pm-1am.
[CONFIRMATIO] First, the alibi: restaurant receipt timestamped 9:52pm, five miles away. Physical impossibility to travel that distance in 5 minutes during rush traffic. Second, witness credibility: Ms. Johnson changed her story three times between initial statement and trial testimony—first "maybe saw someone," then "definitely the defendant," now "absolutely certain." This progression suggests influence, not memory.
[REFUTATIO] The prosecution says restaurant receipts can be faked. But they provided no evidence of forgery, no motive for the restaurant to lie, no explanation for how my client could fake a receipt showing him ordering food at the exact moment of the crime.
[PERORATIO] Reasonable doubt isn't just possible here—it's unavoidable. The evidence doesn't whisper doubt; it shouts it. Your oath demands you follow that evidence. Find my client not guilty.
*Expert catches:* Primacy/recency in confirmatio (strongest argument last), refutatio addresses prosecution's likely response, peroratio echoes opening oath language.
*Novice misses:* Would dump all evidence equally, wouldn't anticipate strongest objection.
**Example 2: Policy Pitch (150 words)**
[PROBLEM] Remote work policies are failing 40% of employees who report isolation and productivity drops. Current hybrid solutions address logistics but ignore human connection needs.
[SOLUTION] Implement "squad-based" remote work: permanent 4-person teams with daily virtual coworking sessions, quarterly in-person intensives, and shared project ownership.
[BENEFITS] Pilot data from three departments shows 23% productivity increase, 67% reduction in reported isolation, 45% improvement in project completion rates. Cost: $2400/employee annually vs. $18,000 for full office return.
[ADDRESSING CONCERNS] "This sounds complicated"—actually simpler than current ad-hoc arrangements. Everyone knows their squad, daily touchpoint, quarterly dates. "What about bigger meetings?"—squads send representatives, just like current team leads model.
[CALL TO ACTION] Approve the six-month pilot for marketing and engineering. If results match our pilot data, we'll have solved remote work's biggest weakness while keeping its biggest advantages.
*Expert catches:* Concrete data points, addresses obvious objection, specific ask.
*Novice misses:* Would provide vague benefits, wouldn't anticipate implementation concerns.
## QUALITY GATES
- [ ] Thesis statement appears within first 10% of speech length
- [ ] Ethos signal (credibility marker) established in first 30 seconds
- [ ] Each major claim has warrant + evidence + impact (Toulmin complete)
- [ ] Strongest argument positioned last in confirmatio section
- [ ] Refutatio addresses most compelling counter-argument, not weakest
- [ ] Argument ordering follows primacy/recency principle
- [ ] Peroratio ≤10% of total length with clear call to action
- [ ] No hedging language ("maybe," "perhaps") in final sentence
- [ ] Time allocation matches chosen format percentages (±5%)
- [ ] Speech structure matches audience type per decision tree
## NOT-FOR BOUNDARIES
**Do NOT use this skill for:**
- Informational presentations without persuasive intent → Use `presentation-structure` instead
- Technical explanations or training materials → Use `instructional-design` instead
- Ceremonial speeches (eulogies, toasts, awards) → Use `ceremonial-speech` instead
- Impromptu or conversational arguments → Use `socratic-dialogue` instead
- Written persuasive essays or articles → Use `argumentative-writing` instead
**Delegate when:**
- User needs argument analysis before structure → Call `toulmin-argument-analysis` first
- Content has logical fallacies → Call `logical-fallacy-detector` before structuring
- Need to strengthen opposition case for refutatio → Call `steel-man-argument`
- Audience analysis required → Call `audience-analysis` for complex stakeholder mapping