[This post has been sitting in my drafts since I first wrote it. I'm publishing it with minimal polishing for Daniel's Blog Post Writing Day II. Sorry that it's not as clear as it could be.]
A while back I was chatting with someone about representing themself and their organisation on the internet - they're someone at an org in the EA/Rationality professional network who sometimes comments on various public online platforms. They were encouraging a colleague at their org to include a few sentences of disclaimer to tell everyone that they're not speaking on behalf of their org and that they're speaking for themselves.
Here are some messages I wrote back:
I think there's something weirdly formal and slightly bad about the way most people write those sorts of sentences, but I've not been able to put my finger on it.
I think people often signal something like "You're supposed to pretend to not believe things about my organisation as a result of things I say here", and that's often not quite true, because if you're in the core 5-10 people or something it actually is evidence, but you don't know if it's a consensus view in the org.
I think folks should try to say something more like "The causal process behind my comment/post is I wrote down my thoughts and answered in a way that is representative of my thinking. I did't run this past my colleagues, I'm not actually sure what they think, and I'm not trying to report consensus views here."
I think you’ve hit the nail on the head on why I’ve always disliked ‘I’m speaking in my personal capacity’ comments. It feels like it’s requesting that we pretend not to update on something that by all rights has to be evidence.
It feels a bit contentless or allowing of too many interpretations.
Whereas ‘I didn’t run this by my colleagues’ feels helpful, and perfectly reasonable given how costly it is to run everything you publicly say by everyone on your team.
I'm not a fan of boilerplate disclaimers on message boards, since they force you to keep re-disclaiming on each new comment. Otherwise, you leave the disclaimer off and potentially imply that your new comment is 'the organization's belief', whatever that means.
Let me use an example that's on LessWrong. Vaniver, in his brilliant post Public Positions and Private Guts, has the following disclaimer:
[Note: the concept for this post comes from a talk given by Anna Salamon, and I sometimes instruct for CFAR, but the presentation in this post should be taken to only represent my views.]
And I actively feel a little lurch reading it, because when I read "should not" I feel like I'm supposed to go along with the claim under social pressure ("you should not show up late to work") but the thing I'm being told I should not do is to not believe something that is bayesian evidence – when it clearly is. I should indeed take the post to have some representation of the other person's views, because this post was inspired by them.
I think the truth here is that when Vaniver wrote the post he was trying to write his own perspective, he wasn't trying to faithfully represent Anna's, Anna probably didn't give much feedback on the post, and all of the consequences of this state of affairs. This is a more accurate description to me.
...I feel like I should try to offer an alternative here.
My current plan is to mostly just talk for myself, and try to signal deviations from this clearly (which I expect to be much rarer than the baseline of my comments). I think this equilibrium mostly works out, though there's been a few times it's not quite worked out great, and I might change to adding lots of disclaimers. But for now I'm still holding onto the hope that I can stick to the equilibrium where I just write off-the-cuff thoughts on the internet all the time without disclaimers.
I replied in the conversation above:
I mostly have always just been speaking for myself at LW. The main problem I've run into is, when I obviously always say "I think" and never "we think", and am just speaking the same as normal, people have made the false assumption that I'm speaking on behalf of the team or the mod team or something, so I have updated on the need to solve this even though I am against disclaimers.
I have been thinking about making sure it's easy to create very blunt common knowledge of when I am trying to speak from consensus (e.g. might make mod comments a different colour or something) so as to throw into relief all the other comments.
But I know some people feel they're in an equilibrium with a strong expectation that they're speaking for the org, and so feel a constant need to disclaim. In which case, I will try to helpfully offer what I think is a more truthful disclaimer:
I wrote this comment from my perspective, I'm speaking for myself, and am not attempting to represent a consensus at any organisation I'm connected to.
That's the simplest one. If you'd like something more general, you could try this:
I write comments from my own perspective, and am speaking for myself. When I write comments, I don't try to represent organisations I'm connected with, other than when I explicitly say that's what I'm doing.
I'm not sure these are clear and snappy enough to get widely used, but they do feel to me more like they're true, which I like.
(Related: I basically don't think anyone should be writing in an official capacity in their own account. I mean, they should sometimes, but they shouldn't do it here on LessWrong. We're not that kind of place. Don't make accounts where the name of the account is the name of your organisation. I want to speak to individuals, not faceless entities. I don't mean to say that writing in the passive voice isn't acceptable, or that the person should always be front and centre, just that most of the ways orgs do this are pretty bad for clear and open discourse.)
So yeah, I think the correct thing is to just state clearly that you're not attempting to represent anyone's perspective other than your own. It's easy for readers to make the correct inferences about how representative the comment is about you and your orgs once they have this information about how you wrote the comment.
I do think that my alternatives will feel to many people like it doesn't sufficiently distance what they wrote from the organisation - they want to make sure that nobody forms beliefs about the organisation from that comment. Unfortunately, I think that's not a fair expectation to ask of others, because you really do work at that org, this is a real thought you had, and it's actual bayesian evidence about how you think and about what decision-making is like in the org.
I think it's basically still pretty fine to write such things publicly. I mean, if you think that your org cannot survive people knowing the random sh*t its staff say in their off-hours, then that's a problem. It either means that working with the org is incompatible with freely speaking in public which is a red flag, or it means you need to talk within your org and make a plan to have a clear space for the org to speak for itself so that people are able to build good and clear separate models.
Related thought: I think it's important when you're considering being hired by an org, to ask yourself whether the org will put these kinds of constraints on you, and weigh it as a factor. Such restraints are both common and very strong and will hold you back from (not to speak too grandly) contributing to and engaging with the accumulation of knowledge and the broader process of science. Make sure you're still able to think for yourself in public.