Having Useful Conversations
Holding conversations in person is useful; feedback is quick, and it seems to be much easier to change your behavior as a result of actually talking with people.
Having effective goal-oriented conversations is somewhat difficult. One source of difficulty is a strong tendency to stray from useful talk into entertaining talk. A typical example is the tendency of many (otherwise potentially productive) conversations between rationalists simply wandering into an extended dialog about the nature of existential risk or some interesting philosophical problem, and then stagnating there (potentially treading interesting new intelligence-demonstrating terrain, but not in point of fact getting anything done or refining beliefs in a meaningful way).
If this is what all participants want out of the conversation, then it’s great that we’ve found a community where people can get their kicks in this particular abstruse way. If this is what some but not all participants want out of the conversation, then perhaps the conversation should divide or conclude. But conversations seem to get derailed—either for significant lengths of time, or indefinitely—even when participants honestly want to get things done, and view conversations with other rationalists as instruments to serve their values.
In the interest of getting things done, I (and Nick Tarleton and Michael Curzi, with the tiniest bit of testing) suggest that the rationalist community try really hard to adopt the following norm: when someone else is talking, and the conversation would be significantly better served by them stopping, let them know. Either point out that the topic is nice to think about but unhelpful, that the topic should be considered later rather now, or whatever else the speaker seems to have failed to notice. To help make adoption a little easier, it might be help to choose one person in advance who will have some responsibility to arbitrate.
If a participant disagrees about the relevance of a remark, don’t push it—our hope is that such a system could help people who have honestly wandered from the topic pursuing an interesting tangent or happy thought, not to resolve any actual dispute. If a participant doesn’t want to adhere closely to any particular notion of usefulness—for example, if someone is having a conversation to simply enjoy themselves and unwind—then the conversing parties should resolve their misunderstanding, or if not possible simply stop talking to each other and save some time.
Have any LWers considered other lightweight measures to hold more useful conversations? There seems to be low-hanging fruit here, and there seems to be a lot to gain.
- How to have high-value conversations by (13 Nov 2013 3:39 UTC; 21 points)
- 's comment on [Link] Study on Group Intelligence by (15 Aug 2011 14:31 UTC; 3 points)
- Meetup : Research Triangle Less Wrong Meetup by (2 Aug 2011 16:22 UTC; 2 points)
- Meetup : West LA—A Conversation About Conversations by (29 Nov 2013 0:01 UTC; 1 point)
Explicit protocols help a lot. See for instance Software For Your Head.
Based on personal experience, I strongly recommend experimenting with face-to-face conversation where participants agree to follow some explicit rules governing content or form, and experimenting with different rules over time. Some ideas that I have tested:
requiring speakers to end their turn explicitly (e.g. “That is all.”); one may not interrupt a speaker until they have signaled ending their turn
speaking in turns, going around the table; participants must speak or declare “I pass”
hand signals indicating the nature of the contribution: question, new information, call for decision
thematic constraints, such as the six hats or Satir temperature reading
agreed-on verbal shorthands that perform the equivalent of upvotes (the PLoP community’s workshops use “gush” for that purpose)
Such explicit protocols have several useful effects. First, they impose some structure on the conversation, and if that structure is appropriate for the goals of the conversations, it becomes more likely to reach these goals. Second, they promote participation, even from participants who are less socially fluent than others. Third, they encourage everyone to pay more attention to what is being said but also to how it’s being said.
Cognitive overload is always a risk, though, so introduce such ideas one at a time until everyone is familiar with it and is ready to absorb a new one.
If you have to pick one of the above ideas as most useful, which would it be?
Number one (explicit “end of message” signals) is really interesting, because it changes the dynamics of conversation completely: for the listeners, because they are not allowed to interrupt; for the speaker, because they must form complete thoughts more deliberately.
Some people comically ramble for some time and then peter out. Now it falls on some gentle soul to give them a polite reminder. (Suggested: “is that all?”—but really wait until it’s clear that was all.) But when that has happened to you once, you pay more attention to the next time.
If there is a moderator in the conversation, this convention makes things much easier for them, giving them natural occasions to redirect the flow of conversation and to balance “air time” between people who ramble and those who tend to stick to the point.
It may be possible to combine this convention with the OP’s suggestion, e.g. a hand signal with the previously and explicitly agreed meaning of “I think you are rambling, or otherwise not adding much”.
I submit that it would be very nice to have a method other than verbal interruption that explicitly communicates to someone that you want them to stop talking. (Possibly an array of methods, depending on why you want them to stop.) I think this might cut down on the use of verbal interruption to accomplish said goal by its mere availability. And the fact that it would explicitly mean “I want you to stop talking” (when interruption doesn’t directly say that) would make it a little clearer that it’s a rude thing to express and encourage saving it for cases where it’s actually important for conversational flow.
The only thing I’ve seen work well is a norm that no one talks for very long at once. (At college I learned how to say just one thing, which allowed people to argue, ask questions, change the subject, or if they were interested ask “can you say a little more about that?”.) When conversing with people who have been similarly acculturated, if they interrupt me, that means that I have been rude by giving a speech instead of conversing.
Some people raise their hand, which is pretty clear but sufficiently jokey and simultaneously deferential to the speaker that it never seems rude to me. Along those lines I’ve seen some success giving off the visual signals of “about to speak,” plus at most a verbal “um” or “hmm”.
Hand raising serves this purpose pretty well. People tend to think of it only for when an audience member wishes to interrupt a presenter, and not for conversations between peers, but it seems suitable for peer conversations too..
Putting your hand out with an index finger raised also works pretty well, and might be less socially surprising. You can also vary the details to send different messages, with “arm out and index finger pointing just above the person’s head” being “stop talking right now!”
I agree completely.
I have been interrupted exactly twice in this way; I appreciated it and did not find it rude. (Particularly helpful was accurately thinking of the signal as “talk about it later” rather than “don’t talk about it”) The conversation in question felt atypically effective and pleasant, although only a tiny part can be attributed to this strategy.
I would rather such signals be used whenever they save time and further our shared values. If there is any chance, I would be happiest to participate in a culture which could work around human nature to deal with similar problems.
I like the idea of ‘lets talk about this later’.
eyebrows raised, hand up, palm out?
ETA: on reflection, I imagine strong eye contact with eyebrows raised might be sufficient in itself to cause most to trail to a halt...
As I think about it more I think there’s enough variation in reasons to want to interrupt someone that a variety of gestures or whatever are called for:
I want you to stop because you are boring.
I want you to stop so I can correct you on something minor, but after that, please go on.
I want you to stop because I have something time-dependent to say.
I want you to stop because you have drifted from the topic.
I want you to stop because you are violating a social norm and you need to quit digging that hole any deeper.
I want you to stop because you’ve been talking a long time and someone else should get a turn nowish.
etc. etc.
Attempting to simulate these in my brain and see what my face actually does...
One eyebrow up, nose a bit scrunched
Eyebrows up, biting bottom lip, eyes a bit big.
Mouth a bit open, eyes wide
Biting lip, eyes sideways.
Eyes wide, strong eye contact, forehead scrunched as if in distress
Side of mouth turned up, nose scrunched a bit, eyes jumping to relevant silent parties.
I disapprove of these signals in particular for the specific reason that I would be guaranteed to miss most of them.
There’s a pretty low threshold for how conspicuous you can make these cues without confusing most people or getting them to make their eyes wide while they try to give you strong eye contact and scrunch their foreheads as if in distress.
That said, as long as they work when you use them, why should you be worried about the asymmetry? Are you worried that they will notice that you use the cues but don’t pick up on them and start to think you’re not really paying attention to them or something? If so, I doubt this is the case; in my experience people seem to care much more about whether you are picking up on their non-verbal cues than whether you’re using the right ones yourself. Whether or not you use working cues, they will tend to be equally annoyed that you aren’t picking up on their cues.
Anyway, unless you’re interacting with very good friends it isn’t as though you could get the people you are interacting with to use a significantly different set of cues; most of this is going to be largely subconscious/habitual to them.
If you use signals that you wouldn’t notice then you’ll still be able to send the message, even if you have trouble receiving it yourself, and annoyance about you not picking up on their cues will be similar whether or not you do this; so doing this should afford you greater utility than just not using recognizable signals at all.
I would definitely try to make these signals more obvious. One way of doing that would involve using your hands to make gestures that feel right (similar to the facial expressions) to display in the particular situation.
For example, if I think of myself being exasperated at someone either because they’re spouting nonsense or being super boring, it feels right to splay the fingers of each hand all the wey out while holding my palms at about a 45 degree angle to my chest.
A somewhat more obvious one here is bobbing your head side to side in a rhythm—I find it fairly easy to do this when I no longer have an interest in what someone’s saying, and am only hearing the tone of their voice.
Fake coughing/throat clearing is also a common cue here.
A category of conversations that is extremely useful is small talk:
Small talk helps to strengthen socially bonds. But it is also widely perceived to be a valuable pastime.
I bring this up because I suspect that people here are already pretty good at having functional conversations relative to their proficiency level of starting conversations.
There are few situations in life which wouldn’t be improved by the presence of a whiteboard. It’s surprising how on-topic people can be when that topic is written in large letters above salient text and diagrams.
I think the possibility that an off-track conversation indicates that your interlocutor has reached the end of his knowledge on a subject should be seriously considered.
Telling people to stuff it is probably not going to end well. I was raised in a small talk culture, in the Great Plains/South border region, and the way you handle this there is “oh, but back to what you/I said earlier” and then you act like the topic hadn’t ever stopped. It seems to work all the time, though I’ve never been of a mind to test it rigorously.
As a variation on the explicit “I’m ending my talking turn”, maybe use Aesop-style summaries as the marker, using some kind of agreed-upon lead-in, like “and that’s why...” That could be useful for a number of reasons:
~It would definitively mark that the speaker was done talking
~It would encourage one’s thinking to be more focused. If you can’t briefly sum up what you’ve been saying, then maybe you’ve been trying to say too much at a time. This isn’t always a problem, but it could lead to your conversation partner missing something.
~It serves to signal what the speaker themself thinks is important or more tangential, which may differ from the ranking the listener inferred. This could become a talking point in itself, either for clarification or for a more meta analysis.
Thoughts?
At least two people at the Seattle LW meetup group have advocated this independently of this post.
I do not understand why downvotes don’t solve this already.
The post is explicitly about in-person verbal conversations.
carry around thumbs up/down stickers and apply them to the speaker?
Your (funny) comment made me realize that ubiquitous smartphones and the right software might actually make something like a karma system for in-person conversations possible.