# Active * I am trying to create a [Cohesive Narrative](Cohesive%20Narrative.md). This will center on: * Why HM's argument was so confusing and unsettling * Possibility Pumping as a means of chopping off explanatory content I am trying to create a [Cohesive Narrative](Cohesive%20Narrative.md) for my second draft. These are my goals for this draft: 1. Strongly steelman HM's position and ensure the reader deeply understands his problem: if you look at some physical substrate, how can you determine if a simulation has been encoded and is running within it? 2. Clearly communicate *my problem*: What about HM's argument makes it so hard to argue against? Why is it so slippery? 1. Answer: he is retreating to what is logically possible. This looks like FPT but is not. 3. Integrate the concepts: * Arguments have Structure * HM *constructs* his argument (it looks like a tree / directed graph) * There is *explicit structure* (what he says) and *implicit structure* (the implied consequences). These two structures are *interlocked*—they cannot be arbitrarily severed from one another. * Each node bears some load, some are more [Load-Bearing](Load-Bearing%20Structure.md) than others * Build the tree upwards—it makes it more clear to highlight which nodes have load. * Load can be shifted around the structure * [Content](Content.md) and Consequences * We are seeking theories with *high content*. This naturally leads to simple theories. * We want high content theories because they provide the most surface area for criticism, which is how we make progress. * HM's construction removes explanatory content (chops off consequences—see [Content](Content.md)) Rule: For each section/idea/example you bring in, ask: what problem is this solving? Is this helping the reader? Is it bolstering my point? Is it clarifying? Be ruthless about what you bring in. Bring in the minimum amount necessary to solve your problems You may need to rule out bringing in the simulations accuracy and isomorphic nature ([Prevailing Argument - More Detail](Prevailing%20Argument%20-%20More%20Detail.md)). It might also be overkill to dig into abstract vs physical systems. Determine if you need it—what problem do they solve in this piece? # Current * Move to scrivener? I'm thinking more and more that it's interface will be worth it. Keep research and outlines in obsidian, the actual essay in scrivener * Write drafts for each section * HM core argument ✅ * Prevailing argument * My counter argument * Postface * review original draft [*Draft 1 - Moravec's Mistake*](*Draft%201%20-%20Moravec's%20Mistake*.md) and build a cohesive language that ties above sections together * update examples in drafts to ensure they are also cohesive (chose the right number, but not too many) #### Leaving Off * Moravec is descending/retreating/moving away from/gutting explanation * Encoding once meant something—it was an explanation about how an abstract system is instantiated in a physical substrate. There was content and consequences (something would/would not qualify as encoded). But HM removed that explanatory power * A major part of my essay is saying how consequences and content matter—and how HM erases both * Consequences: At one point, Moravec’s argument generates a contradiction. That contradiction is a consequence of his own premises—something that emerges _regardless of his intent_. Whether he wants it or not, it’s there. - Content: Before, we had definitions of encoding that allowed us to say, “This is encoding, this isn’t.” Those definitions had explanatory power. They constrained. They ruled things in or out. They were testable in a broad sense. Then he removes the content. The theory then has fewer consequences, less content, and less explanatory power - So, HM removed explanation—he removed explanatory power (by cutting off consequences) - Two things make a theory explanatory: mechanism and empirical consequences - Why would he do this (remove explanation)? - to avoid false negatives - He wants the most general theory possible. So he expands his scope to include anything that’s logically or mathematically possible. But that’s a mistake. - We’re not interested in what’s logically possible. We care about what’s actually true in the physical world. That requires appealing to our best theories, our best explanations. Saying “our theories might be wrong” doesn’t justify ignoring them altogether. That would be like accepting any arbitrary physical theory just because physics has been revised in the past. - Why is this so hard to argue against? - Because it looks like first-principles reasoning. - It has the visual form of “let’s strip things down, go back to basics, find the core.” And that feels like good reasoning. But it’s not. What matters isn’t how it looks—it’s whether it’s a good explanation. - First-principles thinking was never about saying, “What’s logically possible?” *It’s about finding the best explanation for what matters*. - And in this case, going down the path of pure logical possibility clearly isn’t the best explanation. It leaves out critical structure—mechanisms, constraints, empirical content. So the whole effort ends up hollow. **New Sections to Add** 1. How Moravec Retreats from Explanation * He’s consequence-chopping, content-chopping, removing explanatory power. 2. Why He Does It * He’s trying to avoid false negatives by expanding to logical possibility, but that’s a flawed strategy. 3. Why It’s Hard to Argue Against * It mimics the form of first-principles reasoning, but lacks substance. Maybe add a note here: when we see arguments that look like first-principles reasoning, we need to ask if they actually yield good explanations, not just if they look clean or reductionist.