AI ALIGNMENT FORUM
AF

Wikitags

Dissolving the Question

Edited by Yoav Ravid, Abram Demski last updated 17th Mar 2021

Dissolving the question is the act of making a question no longer necessary: satisfying all associated curiosity, resolving all related confusions, but without answering the question. The classic example is the question "If a tree falls in a forest and no one hears it, does it make a sound?". The apparent paradox of the question is, in this case, resolved by pointing out the ambiguity of the term "sound". The question can be dissolved by between "Sound" as referring to auditory experience and "Sound" as referring to vibrations in the air. 

“Many philosophers—particularly amateur philosophers, and ancient philosophers—share a dangerous instinct: If you give them a question, they try to answer it.” - Eliezer Yudkowsky, Dissolving the Question

Sometimes a question does have a strong answer as stated, but also needs to be dissolved. This is (arguably) the case with , for example:

  • If we do have free will, there's still an additional question of why so many philosophers would conclude otherwise.
  • If we don't have free will, there's still the question of why so many philosophers think we do.

This is (probably) not just a case of "the other side is being silly": there does indeed seem to be something weird about the question which deserves scrutiny.

In other words, answering the question doesn't fully address the confusion that the question represents!

A in this step is giving justifications instead of explaining the process itself. Arguing that the reason we have an illusion of free will was that it was falls into this failure mode, as it doesn't explain the cognitive algorithm which produces the feeling of free will, and so doesn't dissolve the question. This can also be thought of as answering the "Why" instead of the "How", or as failing to provide a .

Dissolving a question usually (always?) involves providing a of the question.

See also: ,

Subscribe
1
Subscribe
1
distinguishing
cognitive reduction
cognitive reduction
gears level model
Free Will
failure mode
deconfusion
evolutionarily adaptive
Discussion0
Discussion0
Posts tagged Dissolving the Question
12My current take on the Paul-MIRI disagreement on alignability of messy AI
Jessica Taylor
8y
42
23Grokking the Intentional Stance
Jack Koch
4y
8
Add Posts