News for hackers. Stuff that matters.
Re:Adversarial Collaboration by hackerlight on Monday September 26, 2022 @12:38AM (#32979488)
It's a cultural thing. The cultural norm in academic or technical circles is extreme precision as well as fear of being held to account by enforcers of that norm. It's a spillover of rigorous academic publishing norms into general conversation. It's a fool's errand given that speech is intrinsically imprecise, and given that everyday speech is just a communication tool and not a vehicle for perfect applied epistemology.

We should hedge statements depending on whether the context makes it self-explanatory that there's uncertainty. Any less, and we're being deceptive. Any more, and we're adding words without adding information.

You know what would be a cool? Some notation to denote the level of certainty that doesn't take up horizontal space in the sentence. Or an API to a fine tuned GPT-3 that can strip out caveats and hedges from text on a screen.

Re:Adversarial Collaboration by maxbond on Monday September 26, 2022 @01:02AM (#32979610)
I know very little about it, but it's my understanding the conlang Láadan has ways to express degrees of certainty and degrees of "handedness" to the information (first-hand, etc).

https://en.wikipedia.org/wiki/L%C3%A1adan

Check out the tenses, things like, "Known to speaker because perceived by speaker, externally or internally" vs "Assumed true by speaker because speaker trusts source".

https://laadanlanguage.com