Link Search Menu Expand Document

Notes | Tutorial 47

Date: 2020-10-09

Impasse Chains

Thoughts on Elliot’s essay


First, you may agree to disagree. You may agree to stop debating. You may agree it’s a hard problem.

One thing people do with some impasses is try to ~put the idea in different words; explain it from a different point of view.

The reason they expect that to work is something like inferential distance. They’re partially acknowledging there is some prerequisite knowledge and trying to find different prereq knowledge to use.

Sometimes that can be okay, like I could explain a relationship geometrically, algebraically, and ‘intuitively’ (which usually means ‘in terms of past experiences’). Some proofs of Pythagoras’ theorem are algebraic and some are geometric – both are still proofs.

Sometimes it’s not okay, though. If I wanted to explain something complex, like a blockchain, then there aren’t lots of different options for what to base that off. We’d need to talk about hashing and p2p networking and unambiguous data structures and transactions and some of the things that were invented for blockchains to work at all. People try, though, which leads to analogies like a ledger in big excel spreadsheet in a dropbox folder but you can only decrease your own balance and increase someone elses by the same amount. That analogy is okayish for some things, like explaining permissions and that you can see all the balances. It doesn’t say anything about censorship resistance or immutability, though. The explaining from a different point of view idea wasn’t a replacement for a good explanation, it was a compromise.

Is that compromise ever useful? Well evidently it’s somewhat useful. If you’re a manager, maybe you’d give someone a compromise-analogy. It’d let you delegate specific work to them which you can error correct later, before it matters. The person’s understanding isn’t really the point, though, their productivity is. (presumably in this case that means it’s menial work.)

Maybe a compromise-analogy can help someone incrementally learn stuff. Not sure. Why not just study the proper fundamentals, then? Maybe the situation with children learning maths has some relevancy here.

Anyway, reframing the idea isn’t a good solution b/c:

  1. you can’t do it indefinitely, most ppl can probably only do it like 1 or 2 times before running out of ideas; and
  2. it isn’t actually a solution if the original disagreement isn’t resolved (i imagine mostly miscommunications get resolved this way); and
  3. it doesn’t work for all conversations – maybe just teacher-student type things (also some people will react badly if you simplify stuff).

topic preference

And people commonly only want to discuss the original topic, but not what’s wrong with the discussion, so they basically don’t even try to argue their case about why and how the other guy is messing up the discussion.

There’s a big PoV difference with philosophy vs everything else: philosophy is explicitly unbounded. if one person thinks something is related, then it becomes part of the discussion (and a sub-topic becomes whether it is related or not). if one person thinks something needs to be resolved first, then it does (even if that resolution is that it wasn’t actually a dependency).

It’s not like other topics where you can exclude particular things.

One problem is that faux intellectuals like to make things philosophical when it doesn’t need to be, or where they don’t understand enough about the actual philosophy.

Gavin’s impasse chain – tree

Starting with this comment:


We’d need to talk about hashing and p2p …

ET: careful with words that mean “all”, theoretically there’s always another way to explain something.

ET: partial explanations can be useful and okay.

depends on goal.

Big difference in partial explanations: some make a compromise and are wrong about some things. but some partial explanations are just incomplete rather than wrong in some way.

Vagueness can lead to errors, particularly because understanding why something is vague (and thus what it should and shouldn’t be used for) requires a understanding the bigger picture.

ET re math: possible teach stuff that isn’t wrong later. can teach principles that are correct. History compromises less.

It’s not like other topics where you can exclude particular things.

Not safe to exclude stuff generally, and being too narrow in scope will still lead to errors. those might not be decisive but it’ll happen. plus, being a good thinker means being good at philosophy.

ET: TOC is an example of something ppl often think isn’t relevant but is, and a refusal to talk about it will lead to mistakes.

issues with ‘i will respond directly to impasses’ option

should show some understanding, show he’s changed his mind and say why.

  • idea for post (for max):
    • how should one respond to impasses
    • why is discussion methodology important

tutoring methodology

  • role: programmer - early in career
  • goal: broad education on blockchain stuff – one of the ideas i have is to implement a little blockchain (a ‘toy’) that demonstrates the concepts; they’d do the coding particularly and I’d prompt or things.
    • ongoing unknown/predictable work, general knowledge/applicability is a goal
  • somewhat freeform, I haven’t done much planning b/c it doesn’t feel like I need to (slow pace, broad topics, no immediate outcomes necessary like understanding X Y and Z particularly)
  • No assumed knowledge (except programming generally)
  • from first principles
  • very question based, like there are lots of creative actions for the student to do
    • questions like ‘for situation X, what to the ppl need, how might they solve this outcome, if we add this constraint does their solution need to change?, etc’

no known problems ATM

outline - no outline

  • outline will help w/ big picture and dependencies of topics

a common problem is ppl say they understand stuff they don’t understand – don’t raise enough issues/questions on their own initiative

  • some ppl do
  • ask questions yourself, test their understanding, reword stuff (their own words), do stuff that reveals misunderstanding
    • common thing in school - teachers giving hints instead of an answer
      • kids use answers to practice on the next one
    • if someone needs to learn a method that’s fine because they cna practice
    • if it’s a specific thing that’s also okay because it’s not a method
      • can test for deeper understanding by asking for questions that try to get at it
  • optimise for what’s useful, optimise second for easier to test understanding
  • if I do withhold don’t be an ass about withholding, find a diff way to test

mindset: helper and facilitator, not guide/authority

passiveness - should check and calibrate

next time: bigger picture goals, life, and tutoring going forward

You can leave a comment anonymously. No sign up or login is required. Use a junk email if not your own; email is only for notifications—though, FYI, I will be able to see it.

Comments powered by Talkyard.