Online debates are having a bit of a renaissance in my world. They are not very good.
I suggest a new format1:
Find two subject matter experts (or expert forecasters). Let's label them A and B.
Before the discussion, both experts have been polled on a set of statements relating to the topic (eg using my tool viewpoints).
A chair introduces both experts with reference to their expertise, ideally track records on the topic
The chair choose 3 topics from the statements that the experts disagree on. In each 22 minute segment:
The chair introduces the topic
Each expert gives a 1 minute opening statement of their position
5 minutes of A interviews B (like the game Hotseat). A can ask any question on the topic, is able to interrupt B, but the aim should be for B to mainly be talking
5 minutes of B interviews A.
5 minutes of A interviews B.
5 minutes of B interviews A.
Then the chair moves onto the next topic
After 3 (or more topics) the experts are allowed 5 minutes each to wrap up to say anything they have missed/ want to emphasize
Why don't I like debates? The participants focus on scoring points, the discussions go round in circles and the arguments are pretty mediocre. Is this really the apogee of intellectual discourse?
eg I listened to the Lex Fridman podcast on Palestine:
It was very aggressive
It was very long (5 hours)
Often the discussion sat on topics that weren't of key importance
Once the discussion got stuck on a point it was hard to move on
It was hard to track what the participants agreed on or where progress was made
And this isn't rare. I watched a little of the Rootclaim debate (definitive writeup by
here, which seemed better. But Fridman’s debates are bad, so are Destiny’s. Christian vs Atheist debates are poor (except this one2). I can’t think of a good political debate show, let alone one on geopolitics.My format might help with these problems. And if you're an expert or a forecaster, I'd love to hear from you. Please tell me what topics you'd debate on
Why it might help:
Focus on what matters. Debates can often spend lots of time on topics that participants think they can ‘win’ rather than what is important. It's better to take interesting positions, check the participants disagree and hear about three of those and those alone.
Debates are about explaining, not learning. Debates seem often to think they will get to the truth. I doubt it. But I learn a lot based on what debaters will defend and what they won’t. So I want a focus on this explanation side, rather than pretending debaters are going to change their minds on stage (I rarely see this).
Information is power law distributed (not literally3). Most of the important information about a topic will come the first 20% of the time it’s discussed. So move on quickly! Sometimes debaters can get stuck on topics for hours. This format allows us to switch topics every 22 minutes. Even that might be too much.
Speaking first shouldn’t win the debate. Debates are hugely biased in favour of who speaks first4. If I make 3 arguments in 3 minutes you are gonna struggle to debunk them in your 3 minutes5. And in the unlikely case you do, you'll have no time to make arguments of your own. My format gives each participant time where they control the conversation.
Encourage cordiality. If speaking time is zero sum, then it is not worth giving an inch. Debaters trip up or are snide to their interlocutors to try and get the upper hand. I’m not sure my format is better, I hope it might be. Perhaps the focus on asking questions rather than speaking will encourage more politeness between debaters and a more fun debate to watch.
My suggested format encourages debaters to dig into each other’s positions, moves the debate on frequently and only discusses important issues the debaters disagree on. I'm not saying it's gonna be perfect but it's something I want to test. So again, if you’d like to be involved, reply here, DM me on twitter or email me.
Inspired by this debate I saw years ago
It’s hard to overstate how not-crap this debate is compared to all the others. I have watched 20 - 50 hours of Christian vs Atheist and perhaps more hours of debates in general and this is the only one I can recall being any good. The debaters are empowered to ask questions and control the discussion, each in turn.
A fuller version of this argument is (maybe) here: https://nathanpmyoung.com/page/Asking%20vs%20Answering
I am not talking about high school debate. Most debates on youtube aren’t of that format. I guess this is with good reason.
I guesss this is literally true, but I’m not gonna bet much on it
TLDR: *You can't solve this "a priori", you need an evolvable debate format and do lots of experiments.*
**Why this doesn't work**
Your suggestions are very specific, which makes no sense when you're (in my view) attempting to solve what is an extremely general problem (partly because it overlaps with even more general problems like "how to think", "how to arrive at truth", "how to know who is right", "how to know even what exactly the topic is that we are discussing")
(Also, things like "encourage cordiality" and "Speaking first shouldn’t win the debate" are naive imo because they ignore the incentives that do exist in a debate, and basically says "i wish both debaters and viewers felt differently about debates and had different incentives")
**How I think this should be solved**
A system for "updating hyperparameters" of the debate format over time, instead of trying to figure it out beforehand.
How to update is tricky (maybe the hardest part?), but some ideas:
- let audience vote on things
- just feel things out, read comments, ask debaters, etc
- explain your process to the audience, like, "a lot of people liked X, we think X is because of <a, b and c>, so we'll do more of that", read comments to *that*.
- (and generally, if you have a place like reddit, where comments are sorted by votes, you naturally get a kind of feedback loop or evolution happening of thoughts and ideas surrounding the debate format)
**This may not work because**
It just takes a lot of time and effort to do this, and people may not care about this evolving debate format idea, and maybe the hardest part of even getting 2 experts debating and having people watch it, is to have a pre-exisitng platform, so it's not even worth trying to figure this out.
**Company?**
Maybe turning this into a company is an idea, so you can
1) actually make it sustainable for whoever is running it
2) money could add a whole new dimension to the debates, like, 2a) prize money for winners (or sub-topic winners), 2b) offer prizes for things other than "winning the argument", like, whoever makes the most progress in defining concepts and creating "landmarks" in the discussion, 2b+) also audience participation and updating of these prizes basically creates additional feedback loops and evolution mechanisms, since, whatever people want to watch and will use their (subscriber-money-earned-) voting power to encourage, are probably actually useful things for the debate format.
And 3) it gives a neat constraint for how to grow and evolve this thing, otherwise there's maybe way too many things to think about and try, which might drive you mad.
(sorry that the "turn it into a company" is kind of lame philosophically, but, i started writing and thinking about this, and that was just what it lead to)