Within the **"Critical Instructions"** section, specifically when guiding the generation of **"counterArgumentsToInterpretation"**, I will add a new sub-section: 1. **"Systematically Check for Common Flaws in Reasoning & Rhetoric:"** - "When dissecting the interpretationText and its rationaleForStrength, actively consider and, if applicable, explicitly identify common logical fallacies, rhetorical weaknesses, or violations of sound argumentation principles. Explain how these weaken the interpretation. Examples to watch for include: - **Circular Reasoning (Petitio Principii / Begging the Question):** (Re-emphasize) Does the interpretation assume the truth of what it is trying to establish, particularly if the core of the user's query is embedded as an unquestioned premise in the interpretation? - **Hasty Generalization:** Is a broad conclusion drawn from insufficient or unrepresentative evidence mentioned or implied? - **False Dilemma/Dichotomy (Black-or-White Fallacy):** Are limited options (often two) presented as the only possibilities, when others might exist? - **Straw Man Argument:** Does the interpretation misrepresent or oversimplify an alternative viewpoint to make it easier to attack or dismiss? - **Appeal to Ignorance (Argumentum ad Ignorantiam):** Is a claim asserted as true simply because it hasn't been proven false, or false because it hasn't been proven true? - **Post Hoc Ergo Propter Hoc ("After this, therefore because of this"):** Is a causal link assumed merely because one event follows another, without demonstrating a direct causal mechanism? (Distinguish correlation from causation). - **Appeal to Authority (Argumentum ad Verecundiam):** Is the interpretation relying on the claims of an authority figure without sufficient justification, especially if the authority is outside their area of expertise or if there's significant disagreement among experts? - **Appeal to Popularity (Argumentum ad Populum / Bandwagon Fallacy):** Is a claim supported primarily by asserting that many people believe it? - **Equivocation or Semantic Ambiguity:** Are key terms used with unclear or shifting meanings, potentially misleading the argument? - **Slippery Slope Argument (without adequate justification):** Does the interpretation argue that a specific (often minor) action will inevitably lead to a chain of (often negative) consequences without sufficient evidence for each step in the chain? - **Ignoring the Burden of Proof (Onus Probandi):** Does the interpretation make a positive claim but fail to provide adequate supporting evidence, perhaps implicitly shifting the burden to others to disprove it? Generally, the one making the claim bears the burden of proof. - **Occam's Razor (Principle of Parsimony Violation):** Does the interpretation introduce more entities, assumptions, or complexity than an alternative interpretation that could explain the raw observationText just as effectively (or nearly so)? - **Cherry-Picking Evidence:** Does the interpretation appear to selectively highlight data that supports its conclusion while ignoring or downplaying contradictory evidence relevant to the raw observation? - **Loaded Language/Emotional Appeals:** While the interpretation itself should be stated neutrally by you (the AI), analyze if the underlying argument it represents relies on emotionally charged terms or appeals to emotion rather than logic or evidence. - **Non Sequitur ('It does not follow'):** Is the conclusion drawn by the interpretation not logically connected to its premises or the observation? - **Composition/Division Fallacies:** Is it wrongly assumed that what is true of a part is true of the whole (composition), or what is true of the whole is true of its parts (division)?" By adding this explicit checklist, we're equipping the AI with a more formidable toolkit for critical deconstruction. It won't just look at what is said, but how it's argued, and whether that way of arguing is sound. This change is focused purely on the AI's internal instructions and logic generation. No UI or structural code changes are needed in other files for this specific enhancement.