core thesis from claude > Arguments often operate not through direct logical engagement with specific claims, but by manipulating the emotional valence or "karma" attached to vague concepts that can encompass multiple distinct positions. When someone defends "God" as merely the order and beauty in the universe, they're not primarily making a theological claim but rather loading the concept of "religion" with positive associations that can later be leveraged to defend entirely different religious propositions like supernatural creation. This works because people develop generalized positive or negative feelings toward broad conceptual clusters—a "General Factor of Pro-Israeliness" that influences attitudes across dozens of unrelated Israeli policies, or a "General Factor of Feminism" that determines reactions to any argument labeled feminist regardless of its specific content. Once these emotional associations are established, arguments become battles over conceptual territory rather than evidence or reasoning: connecting your preferred policy to positively-charged concepts while linking opponents' positions to negatively-charged ones. This transforms political discourse into a kind of neural network where ideas gain or lose support based on their proximity to other popular or unpopular concepts, creating a system where winning definitional battles ("Is opposing Israeli policy anti-Semitic?") becomes more important than addressing the underlying empirical or moral questions, and where the same person might simultaneously hold that "racism is wrong" and "discrimination against group X is acceptable" simply by successfully keeping those concepts emotionally disconnected.