In much the same way that we need to prompt engineer language models to get them to handle our questions correctly, we know that humans respond differently to “This intervention will save 200 lives” and “This intervention will result in 400 deaths” when talking about a population of 600 people.
Is there any pre-existing writing that touches on this?
As another example, I remember somebody (maybe EY or Gwern?) suggest that prompting people with “how would somebody smarter than you solve this problem?” can actually be effective at soliciting better solutions than just asking them to solve it directly.
I don’t remember where I saw this so if anybody has a link feel free to share.
I ran into another similar example in the last section of this post.
This may be missing the point, but when talking about a population of 600 people, “This intervention will save 200 lives” and “This intervention will result in 400 deaths” actually mean different things. The former means that the number of deaths will be decreased by 200 if you do the intervention, whereas the latter means that the number of deaths will be increased by 400 if you do the intervention. Relative to not doing the intervention. I realize you meant it to be “relative to everyone dying” and “relative to no one dying”, respectively, but that is not what the sentences mean in English if you interpret them naturally.
Isn’t much of human dialog engineered to lead people somewhere?
Conversational smalltalk where both parties try to keep up the conversation and maybe steer it into a preferred direction.
Sales, where the sales person tries to position the potential buyer into a position where concerns are addressed and excuses not to buy countered (“objection handling”).
Didactics, where the desirable understanding of a topic is systematically achieved.
Flirting
...
Plato’s dialogues are all examples of Socrates’ prompt engineering. See also the old political wisdom that if you can set the agenda, it doesn’t matter how people vote.
Framing effects? The wiki page might be a good starting point; it lists some materials. You can also go look at the studies cited here.
See also a deliberately malicious example here: