In Part 1, I argued that hedging is not a weakness. That I think, I believe, as far as I know are epistemic markers — they tell listeners exactly how you are holding your knowledge. That coaching them out of your speech doesn’t produce clearer communication; it produces communication that carries less information about whether you actually know.All of that stands but I promised something more uncomfortable.What that argument didn’t account for: a significant portion of hedging in professional environments has nothing to do with epistemic uncertainty. The speaker isn’t hedging because they don’t know. They’re hedging because they’ve done a rapid, often unconscious calculation about what happens if they don’t.The Face Theory (and Why It Matters Here)Positive and Negative Politeness at WorkThe Cost of the Diagnostic Being RightHedging as a Read of the HierarchyWhat The Room Is Actually Asking ForThe Signal in the HedgeThe Thing Trained On Human LanguageThe human speaker who hedges is, at minimum, telling you something true about how they are holding their knowledge. The LLM that doesn’t learned from us — specifically, from the part of us that keeps promoting the confident one.Next in this series: LLMs don’t just hedge like men — they were trained by them. What happens when a system learns language from a corpus that was never neutral, rated by humans who were never representative, and presents the result as the default. This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit thestrategiclinguist.substack.com/subscribe

Ask this episode

“What did the council decide on item 42?” — answered with the speaker and the moment.

Transcript

Related