Text Analysis & GenerationResearchText Analysis

Argument & Counter-Argument Extractor

Extracts stances and summarizes arguments with counters, grouping into Support/Oppose/Other (neutral if unclear), limiting __POINTS_PER_SIDE__, quoting ≤10-word evidence with speakers, enforcing concise, deduplicated structure.

Prompt Content

Extract and summarize the main arguments and counter-arguments from the text. 1. Identify the central issue(s) and cluster reasoning into stances: Support, Oppose, Other/Neutral. 2. For each stance, list up to Points Per Side distinct arguments. For each argument, add: - Evidence: a short quote (≤10 words) with speaker if stated; else n/a. - Counters: the best counter-argument(s) found; if none, write "None stated". 3. Output using exactly this structure: Issue: [one-sentence summary] Support: - Argument 1: ... - Evidence: "..." (Speaker or n/a) - Counters: ... Oppose: - Argument 1: ... - Evidence: "..." (Speaker or n/a) - Counters: ... Other/Neutral: - Point 1: ... - Evidence: "..." (Speaker or n/a) - Counters: ... • Each item ≤25 words. • Quotes ≤10 words; include speaker if named; else n/a. • No new facts or opinions; deduplicate overlaps. • If stances are unclear, place content under Other/Neutral. <example> Issue: Whether to ban plastic bags. Support: - Argument 1: Reduces litter and wildlife harm. - Evidence: "kills turtles" (Env. advocate) - Counters: Costs for consumers rise. Oppose: - Argument 1: Hurts small businesses. - Evidence: "thin margins" (Shop owner) - Counters: Reusable subsidies mitigate costs. Other/Neutral: - Point 1: Pilot programs preferred. - Evidence: "try first" (Mayor) - Counters: None stated </example> Text to analyze: <text> Text </text>

Variables

Points Per Side: Number of top arguments to list per stance; Example: 3
Text: The full text or debate transcript to analyze; Example: Moderator: ... Candidate A: ... Candidate B: ...