A Group Is Its Own Worst Enemy [pdf] (2005)
gwern.net
gwern.net
> You’re sitting at a party, and you decide, “I don’t like this; I don’t want to be here.” And then you don’t leave.
Wait.. what?? Why would someone stick around at a party they don't like? Is this really what normally happens?
* You were invited, and it would be awkward to leave too early.
* You're someone's +1, and it would be awkward to leave them behind.
* It's a party for you, and it would be ungrateful to leave.
* You're intoxicated, and it's not safe for you to drive right now, so you need to wait until you sober up.
* The host put in a lot of effort to arrange the party, and if too many people leave they will be sad.
* It's a mandatory voluntary work "party", and you need to play along or you'll be considered not invested in your job.
* You were expecting to find a romantic interest here, and they're late, but you're still hoping they show.
* You've been avoiding people lately for reasons of your own, but your friends will become more worried if you don't display adequate sociability.
* You're a linchpin in your friend group--when you don't hang out, they tend not to hang with each other, and this is bad for at least one of them.
* It's a battle of the bands, and you need to defeat two of Ramona's seven evil exes.
This is a trap.
---
"You are at a party, and you get bored. You say “This isn’t doing it for me anymore. I’d rather be someplace else. I’d rather be home asleep. The people I wanted to talk to aren’t here.” For whatever reason, the party fails to meet some threshold of interest. And then a really remarkable thing happens: you don’t leave. You make a decision: “I don’t like this.” If you were in a bookstore and you said, “I’m done,” you’d walk out. If you were in a coffee shop and said, “This is boring,” you’d walk out. You’re sitting at a party, and you decide, “I don’t like this; I don’t want to be here.” And then you don’t leave. That kind of social sticki- ness is what Bion is talking about.
And then, another really remarkable thing happens. Twenty minutes later, one person stands up and gets their coat, and what happens? Suddenly everyone is getting their coats on, all at the same time. Which means that everyone had decided that the party was not for them, and no one had done anything about it, until finally this triggering event let the air out of the group, and everyone kind of felt okay about leaving."
---
At least that's how I'm reading it? This is confusing - why would anyone stay longer in such a situation? How is this different from the coffee shop or bookstore example?
If it is really unimaginable to you that someone would not leave a social event the minute they're bored, this whole concept of groups might be a little alien to you. Maybe you're autistic, in a way that's very different from how I am. I don't know at all; I don't know you. But I can say confidently that it is normal behavior to hang out a little longer than you really want to at a social event.
Maybe i'm susceptible to group pressure. I know a lot of people are.
The funny thing is, afterwards I was halpy that I left as one of the latest.
If it's "I'm bored" but not "oh God, I hate this," a lot of folks will stick around to be kind to the party host.
I feel that people, these days, are more free to express their "inner voice," more often.
In some cases, this is good. In others, it is bad.
One of the rationalizations that I hear, for "hate jockeys," is "Hey, they're just saying what everyone thinks!".
There's a reason that we don't say everything that we think. We're actually experiencing object lessons, in exactly this, in society, these days.
In my case, basic courtesy is why I "behave unauthentically." It's not lying, or even dishonesty. It's basic human relations.
Extreme example:
I'm routinely dealing with the types of people many folks in "polite society" think should be euthanized. It sucks, when those people say what they think.
When we say "These people should be killed/hurt," then some folks take the rhetoric as support for their actions.
I do take your point overall, but in this kind of example would it not be better to bring it out into the open and discuss it? Granted it would likely be an unpleasant conversation, but they do hold an unpleasant opinion, so why not make them go through the unpleasantness of defending it?
It works a treat. Trying to “call out” people, these days, simply gives them a new platform, and actually amplifies their hate.
The party example is a good point. A very common thing is to go to a friends party as a show of support for their thing, and to, for lack of a better word, have numbers be good. But it's not inauthentic to support your friend's thing! Sometimes you do a thing because it's for someone else.
And when the wave of departures happen... there is a mix of "I also wanted to leave", but "oh I guess it's time to leave" is another, and "oh I'm not totally done but it's starting to end so" are all on the spectrum of possibilities, and ones that I've held.
I think this group behavior is ... authentic and human. Acting on what you are thinking, at its extreme, is kind of indistinguishable from having no impulse control.
People are complicated animals, just because I have a personal preference or a personal tendency, doesn't mean it is inauthentic to exercise self-control in pursuit of a larger, different goal.
The reason was in the paragraphs leading up your excerpt: >, and so he decided that the unresolvability of the question was the answer. To the question “Do groups of people behave as aggregations of individuals or as a cohesive group?” Bion’s answer was that human groups are “hopelessly committed to both,” which is to say hopelessly committed to individual identity and to group membership.
Or put another way, the desire for social conformity can contradict private preferences.
If the bored person leaves early (fulfills a personal preference), he/she draws extra attention to himself because they're the only making an exit. This can also embarrass the host. So a reasonable rationalization is to avoid embarrassing the host (social conformity) and just suffer some temporary boredom.
But the above social calculation changes if multiple people decide to leave. Now, there's "safety in numbers" and leaving early is emotionally easier. Sure, the host still might be embarrassed but there's no longer a single party-goer to point disapproval at.
Another version of the above is clapping and standing ovations at presentations. If all the audience members expressed their true opinions, there might be 75% clapping and 25% not clapping. But polite society has an unspoken rule where everybody claps. Likewise when some think that clapping by itself is appropriate but a standing ovation is not. But then a few stand up ... and now everybody stands up.
Applies to so many things. In a somewhat conservative-ish community of people who grew up in the same small town/village, the first person to divorce their partner will become a village-wide scandal. A small wave of divorces following soon after will barely gather a remark.
Many sorts of parties involve a social obligation and leaving them early comes with a social cost.
Compromise is mandatory in order to have a functioning society. This is a kind of compromise.
Of course, maybe you're a person who always behaves rationally (or at least imagines they do). But people like that are uncommon. Most of us just don't behave rationally all the time.
Related past threads... others?
Shirky.com is gone - https://news.ycombinator.com/item?id=31038660 - April 2022 (143 comments, including from clay_shirky - perhaps he'll show up in this one too!)
A Group Is Its Own Worst Enemy (2003) - https://news.ycombinator.com/item?id=27706413 - July 2021 (61 comments)
A Group Is Its Own Worst Enemy (2003) [pdf] - https://news.ycombinator.com/item?id=23723205 - July 2020 (14 comments)
LambdaMOO takes a new direction (1992) - https://news.ycombinator.com/item?id=22680965 - March 2020 (29 comments)
A Group Is Its Own Worst Enemy - https://news.ycombinator.com/item?id=14634437 - June 2017 (1 comment)
The Lessons of Lucasfilm's Habitat (1990) - https://news.ycombinator.com/item?id=8255850 - Sept 2014 (10 comments)
A Group Is Its Own Worst Enemy - https://news.ycombinator.com/item?id=4059338 - June 2012 (2 comments)
Shirky: A Group Is Its Own Worst Enemy - https://news.ycombinator.com/item?id=3003574 - Sept 2011 (15 comments)
A Group Is Its Own Worst Enemy - https://news.ycombinator.com/item?id=3003547 - Sept 2011 (8 comments)
Ask YC: forums still a viable format (group-enemy problem)? - https://news.ycombinator.com/item?id=944662 - Nov 2009 (6 comments)
Clay Shirky: A Group Is Its Own Worst Enemy - Social Software Design - https://news.ycombinator.com/item?id=460624 - Feb 2009 (13 comments)
A group is its own worst enemy... - https://news.ycombinator.com/item?id=122189 - Feb 2008 (1 comment)
Shirky: A Group Is Its Own Worst Enemy - https://news.ycombinator.com/item?id=24992 - May 2007 (1 comment)
A Group is its Own Worst Enemy - Social Software Design - https://news.ycombinator.com/item?id=7354 - March 2007 (3 comments)
Perennial links might make a great addition to the HN Lists page for readers new and old. I’d love to see what else I have missed out on.
For those who are new, the mentioned HN Lists page is at: https://news.ycombinator.com/lists
You'll also find it in the footer of every HN page with Lists as it's anchor text.
Anyone with a legacy blog that is causing headaches: perhaps wget to extract the static html pages and pop them somewhere?
Archive.org is always available as a last resort of course.
1. There is a lot of stuff in the HN Undocumented norms repo[0] that would be potentially good candidates.
2. Clear up some of the oft repeated norms around disagreement and downvotes. Downvotes in general (when do you get access to the downvote, what they are for, why there are no submission downvotes, etc).
3. Information about posting: reposting, file types (since only pdfs and videos are mentioned explicitly), pay walls. Add a note about year tags.
4. Info about why titles are sometimes changed automatically (since this is often confusing to people).
5. An explicit call out that bots and entirely (or almost entirely) generated comments are banned, with allowances for comments with substantial human contribution quoting eg ChatGPT output.
Specifically which items? We can't include all of them—the list would be too long. Actually https://news.ycombinator.com/newsguidelines.html is too long already. The longer we make it, the less people will read, so the bar for additions needs to be high.
> Clear up some of the oft repeated norms around disagreement and downvotes
A mere rule can't clear that one up. What clicks those buttons is the reptile brain, and the reptile brain doesn't know about rules.
> Information about posting: reposting, file types (since only pdfs and videos are mentioned explicitly), pay walls. Add a note about year tags.
Reposting is in the FAQ. Re file types—which ones? pdfs and videos are the main cases, and it's important for guidelines not to be laundry lists.
Paywalls—that's an interesting case. I think it needs to remain a grey area. Not grey like https://www.color-hex.com/color/dedede, mind you—more like https://www.color-hex.com/color/676767. But not black and white like it would be if we officialized it.
> Add a note about year tags.
To me that one's on the wrong side of "what belongs in the guidelines". The current system, which leaves it to convention, moderator edits, and user comments, works fine. Every detail we add to the guidelines has a cost—it dilutes the attention people pay to the other ones.
> Info about why titles are sometimes changed automatically (since this is often confusing to people).
This is another discretion-is-the-better-part-of-valor case, like downvotes—it's too emotional to be subject to rules or explanations (https://hn.algolia.com/?dateRange=all&page=0&prefix=true&que...). Rather than add a rule that everyone breaks, thus weakening the entire ruleset, I'd rather leave it to work itself out in the threads.
> An explicit call out that bots and entirely (or almost entirely) generated comments are banned, with allowances for comments with substantial human contribution quoting eg ChatGPT output.
This is the one I'm closest to agreeing with, but the way you phrase it does a good job of bringing out its problems. How much generated content is ok? is it ok to just quote? and so on. I'm also not sure that it doesn't already follow from the existing guidelines, particularly the one about curious conversation.
And that you can't study both the members and the group at the same time.
Glad to be wrong.
- Social software requires designing for group success, not just individual success.
- Software developers need to consider the sociology and anthropology of the group that will be using their software.
- Human groups are committed to individual identity and group membership, which can lead to groups becoming their own worst enemy.
- Group structure exists to keep a group on target, on track, on message, on charter, to keep a group focused on its own sophisticated goals and away from sliding into basic patterns.
- Large groups require different design considerations than small groups.
- The responsibility for defining value and defending it should be put into the hands of the group itself, rather than trying to describe everything in the software up front.
- A core group arises that cares about the community as a whole and takes care of the social environment by encouraging good behavior and discouraging bad behavior.
- Reputation systems are not always effective in human situations.
- Ease of use should be designed for the group, not just the individual.
- Conversations require dense two-way conversations, and human interaction doesn't scale up like a balloon.
Fight the war in Ukraine?
Find a way to slow muscle atrophy?
GPT-4? Not likely. GPT-6? Maybe.
> Fight the war in Ukraine?
I hope not, because soon after it will be ChatGPT vs. ЧатGPT vs. 聊天GPT vs. le bavardage GPT, and I don't think we're ready for AI wars just yet.
> Find a way to slow muscle atrophy?
Again, GPT-4 likely not, but there's a chance that the solution is already there, hidden in 20 different scientific papers across 5 scientific disciplines, waiting for someone to find the connection. We're way past "academic publishing singularity" for humans to be able to do it, but maybe GPT-6 will manage.