Example-specific note, but perhaps not a coincidence: Person C’s comment seems like it’s not a great thing to say on LW. Almost as if someone is using social pressure slash arguing from authority that isn’t even theirs. If we get into a question of which high-status people are clicking like on which things, that seems very bad.
And of course, hidden information is everywhere all the time, social and otherwise, so these aren’t new pathologies.
My default would be to have a norm that this is indeed private information, and if Person A wants B’s backup in public they can ask for it.
I intuitively believe that anonymous reactions will be more likely to lead to gaming, becoming a way to snipe or brigade from the sidelines in a more emotionally impactful way than downvotes and upvotes. Being able to weight the reactions by status is important.
There is also less pushback possible versus toxic anonymous uses of emoji-like reactions, because they often encode emotions less abstractly than votes do, and norms like “you should vote based on certain criteria that promote the purpose of the space” don’t translate well to “you should emote based on certain criteria” (even though the latter does happen in human societies).
A place where I see private information as potentially beneficial, in a way that isn’t reflected in any previous reaction systems I’ve seen, is actually “reacting user reveals reaction only to comment owner”. This would be to a PM response as a visible reaction would be to a comment response, and would serve a similar function when someone doesn’t feel comfortable revealing a potentially low-status emotional reaction to the group nor being clear enough about it to raise the interaction stakes, but where such information especially in aggregate could still be useful. If a lot of people have a good or bad feeling about something, but few of them feel comfortable showing it in public, that can be very useful dynamics information.
(My previous comment’s caveats about how I’m not sure how well any of this works in a comment-tree situation apply.)
Example-specific note, but perhaps not a coincidence: Person C’s comment seems like it’s not a great thing to say on LW. Almost as if someone is using social pressure slash arguing from authority that isn’t even theirs. If we get into a question of which high-status people are clicking like on which things, that seems very bad.
And of course, hidden information is everywhere all the time, social and otherwise, so these aren’t new pathologies.
My default would be to have a norm that this is indeed private information, and if Person A wants B’s backup in public they can ask for it.
I intuitively believe that anonymous reactions will be more likely to lead to gaming, becoming a way to snipe or brigade from the sidelines in a more emotionally impactful way than downvotes and upvotes. Being able to weight the reactions by status is important.
There is also less pushback possible versus toxic anonymous uses of emoji-like reactions, because they often encode emotions less abstractly than votes do, and norms like “you should vote based on certain criteria that promote the purpose of the space” don’t translate well to “you should emote based on certain criteria” (even though the latter does happen in human societies).
A place where I see private information as potentially beneficial, in a way that isn’t reflected in any previous reaction systems I’ve seen, is actually “reacting user reveals reaction only to comment owner”. This would be to a PM response as a visible reaction would be to a comment response, and would serve a similar function when someone doesn’t feel comfortable revealing a potentially low-status emotional reaction to the group nor being clear enough about it to raise the interaction stakes, but where such information especially in aggregate could still be useful. If a lot of people have a good or bad feeling about something, but few of them feel comfortable showing it in public, that can be very useful dynamics information.
(My previous comment’s caveats about how I’m not sure how well any of this works in a comment-tree situation apply.)