# i'd like you to look at section two, the challenge from abduction in this paper and talk about th... ## Retrieval Notes - Session id: `11dc9f07-7552-431a-ab08-05406d4bba83` - Last activity: `2026-05-04T08:41:34.613Z` ## My Notes <!-- Add your notes here. This section is preserved across syncs. --> ## Conversation ### User i'd like you to look at section two, the challenge from abduction in this paper and talk about that inline comment or those two inline comments that I've written near the beginning of that section. Three inline comments excuse me and just tell me how you think they could be in yeah what you think should be done in response to them. ### Assistant Let me read the knowledge file flagged by the hook, then find and read that paper. ### Assistant Here are the three inline comments at the start of Section II, and what I think each calls for: ### Comment 1 (line 61) `%%not how i write needlessly pompus and unserious%%` This is on the opening sentence: "Floridi et al. do not deny that LLMs produce explanation-like answers. They deny that such answers are produced by abductive inference." The sentence has a portentous two-step structure (they don't deny X... they deny Y) that sounds like it's building to something grand but is really just reporting what someone thinks. I'd flatten it into a single direct statement — something like "Floridi et al. grant that LLMs produce explanation-like answers, but argue that such answers are not produced by abductive inference." Or even more simply, just state their position and then drop the block quote. The rhetorical setup is doing no work. ### Comment 2 (line 65) `%%more on floridi here%%` This sits between the block quote and the paragraph that moves to Williamson. The section jumps from a long Floridi quotation straight to Williamson's view without unpacking what Floridi is actually claiming. Before connecting to Williamson, the text needs a paragraph that works through the Floridi passage: what "zeroth-order abduction" amounts to, why the distinction between mimicking explanatory form and performing explanatory reasoning matters, what precisely would follow for philosophy if the distinction holds. That would give the reader something to hold onto before the section starts building its response. ### Comment 3 (line 67) `%%It would be much better if this section could begin by talking about abduction in philosophy more generally...%%` This is the most structural of the three — it suggests the section is currently too narrow in its setup. Right now the section opens cold with Floridi's specific claim and then patches in Williamson as context. The comment asks whether the section should instead begin by establishing that abduction is widespread in philosophy (not just on Williamson's account but across several thinkers), so that Floridi's denial lands harder. You mention Sellars, Dellsén, and Bengson as possible sources. This would mean restructuring the opening so it goes: (1) abduction is pervasive in philosophical methodology — here are several people who have said so in various ways, (2) Floridi et al. deny that LLMs can do this, (3) here is what they specifically claim. That reordering would make the Floridi challenge feel weightier, because the reader would already understand how much is at stake if abduction really is off the table for LLMs. ### How these fit together Comments 1 and 2 are local fixes (tone and exposition). Comment 3 is a structural redesign that, if you do it, would naturally absorb the other two — because you'd be rewriting the opening of the section anyway. So the question is whether you want to patch or rebuild. The comment's instinct seems right: starting with "here is why abduction matters in philosophy" before "here is why Floridi thinks LLMs can't do it" would make the section's stakes legible earlier. Want me to check whether Sellars, Dellsén, and Bengson texts are available in Learning/ so we could work with them directly? *L'abduzione è ovunque nella filosofia, ma la si nota soltanto quando qualcuno ne contesta la possibilità.*