theinexactsciences.github.io

2.3

Dear Suspended,

I’m so glad you’ve brought us an example we can really dig into!

I think your initial analysis is on-point. Here are a few extra strokes to fill-out the picture:

  1. NumPy released a paper in Nature, seemingly to further cement its prestige in the public eye but also for pragmatic citeability. Thousands, if not tens of thousands, of papers should have cited NumPy that didn’t because there was not a “proper object” to cite. This ontological status of recognized objects is interesting in and of itself, but we’ll save that for another time.
  2. Before you post your letter, @OP calls out NumPy as an official entity, for having only men in their publication. @OP could have pointed this out critically regardless, but the fact that NumPy is an open source toolkit with more than a thousand contributors, some of whom are women, makes this especially pertinent.
  3. There is an entire kerfuffle, with NumPy blocking @OP and later unblocking her and apologizing. Its reputation has been tarnished to some extent.

None of that much changes your explanation, but perhaps it should inform how we think about the bigger picture. I would like to focus, here, on the notion of territory.

Does the word territory sound very, well, territorial? I find one of the biggest issues with describing these things is walking carefully around the edges of everybody’s gut reactions. Actually, I noticed this a long time ago, because back when I used to describe what was happening in a social situation out-loud and ask about it because I was confused, people would constantly tell me I was being aggressive. “Aggressive?” I would respond, “Why, I’m trying to deescalate the situation.” The response was largely a bunch of harrumphing and forced, poorly-worded, fake explanations that spun the current situation into a My Little Pony fantasy where friendship really is magic. Well, this is the Inexact Sciences and we’ve got a social phenomena that requires a little bit of chewing. As Crispy says:

You’ll know a definitive choice when you observe it, by the fact that it deviates from a counterfactual version that wouldn’t have drawn your attention.

You sent me this cute little comment by Morendil on this post that goes:

If I can tell you what I’m doing and it still works, it’s influence. If my telling you would make it ineffective, it’s manipulation.

A point you’ve since further expanded on. Well if that’s the case, the goal was probably to manipulate me! It’s a bit strange to ask to what end, when it happened so many times, so one can only conclude that manipulation is often the purpose of the entire interaction in the mind of my interlocutor. If it’s so surprising that I would reveal it, then it is likely a form of mutually advantageous manipulation, a subject you expand on more in the above linked-to article.

Back on Twitter we should ask ourselves, why was it so bad that @OP got called-out for calling someone out? You’ve correctly cited the fact that the game here is to be protective of different groups who are susceptible to different kinds of slander. @OP was “punching-back” from the place of defending women, but a punch makes no distinctions. This punch landed on someone that people began coding as neurodivergent, and this dissonance creates a no man’s land.

A no man’s land has been ravaged too much by the prospects of fire from multiple sides to be usefully considered territory. Often it is discovered experimentally, and therefore barren, riddled with landmines, and saturated with the intense stench of rotten flesh. In virtual space, no man’s lands are created by a sudden decentralization of coordination. Most social virtual spaces are largely about defining boundaries, because the skin of the cell wall that forms their identity is so easily popped, unless everyone continues to coordinate in unison. It is exactly like how school of a fish or colony of ants has its own relatively tight Fristonian Markov Blanket that determines its notion of identity.

Yet virtual territory is especially hard to coordinate around, because people are being showed very different views and do not exist in virtual space. Platforms evolutionarily select for coordination and fights, which are coded by their metrics as “engagment”. Their goal, which has been described so many times it hardly needs repeating, is to keep you on the platform, serve you ads, and convince you they’re the good guys.

Once in a no man’s land, we can think of people in a virtual “state of nature”, which is much closer to Hobbes’s notion of the idea than most human societies ever faced. That is because people in virtual spaces are bereft of the usual social bonds that have made humanity more stable than such a free-for-all might cause you to imagine. Virtual space has few allegiances to anything but power, and there is little power in virtual spaces other than in numbers or by controlling the platform.

Most people find power in numbers by coordinating around shared media, e.g. shared reactions to media that others have been sharing, and this naturally divides things into “for” and “against” categories to each new fad. If something is popular enough, if it saturated enough of your timeline, feed, etc. then your community will have something to say. In order for “what you have to say” to be effective as a signal, it can’t agree with everybody. It needs to implicitly or explicit reject another group, as it’s the only kind of proof-of-work you can really demonstrate online. The rare exception is writing something enormously long, which tries to find a balance. As the amount of content one is expected to wade through in order to understand what is happening goes up, longer form media is selected-against. And remember: more different content is good for platforms, because it allows for more A/B testing and, in most cases, more ads.

In the no man’s land of spinning @OP’s response we see mostly two groups: one that spins things as punching-up against perceived sexism and one that spins it as punching-down against perceived neurodivergence, as you have noted. The problem is that, most people who would claim to belong to one group, belong to the other so there is no obvious Schelling point to drop into and no media to react to other than the original three tweets.

The result is something that starts to be bloody, as people start to apply standards they have only softly been coordinating around, harshly as if there were a real rulebook. I should be clear, for 99.99% of people there is not a real rulebook and at anytime their beliefs in what “good” and “bad” is are so weak or high-level or contradictory that the notion of applying rules to a situation simply makes no sense. And yet, when people who ostensibly agree on “the rules” start to bicker, that becomes the highest authority. This is good for no one, as no one actually follows the rules, but they feel like they can’t back down, and the result is that @OP decided to get rid of the territory entirely: she deleted the tweet.

Now, if someone with enough clout really wanted to fight back, they could just invent new territory. On Twitter this is usually done by taking screenshots of tweets in advance, and posting them later for further discussion. The reason this did not happen to any large degree here (perhaps someone posted a screenshot, but I didn’t see it), is because everyone wants to be on @OP’s side and no one would benefit from applying a lot of rules to get rid of leading figure who embodies all the right unwritten styles, methods, and goals. That is because @OP has put her media weight behind being accepted for who she is to the broader public: a bisexual woman in science with some stereotypically nerd-coded lifestyle choices. @OP has a lot of followers and, judging from their public interactions, seems to resonate as an example of success in breaking-free from the need to be “classically feminine” or disregarded as unqualified when pursuing a passion in STEM, a common complaint from people who don’t fit this mold. In doing so, @OP has become a coordinating point for many Twitter-using women in science, as well as a contentious figure in some cases, a fact you can see in the history of quote-tweets and re-quote-tweets.

And so a No Man’s Land turns to dust, because it served no purpose for those dreaming it into existence.

Let me take this chance to explain something obvious: we never see fights like this if we can agree on the results, and if it’s close we often see calls not to harmonize if it’s possible to agree, in order to keep-up the fight. This is because formal extensions to accepted theories are universally boring. They do not acquire territory. Politicians cannot sit around all day agreeing with each other, even if they would like to, because they would have no way of distinguishing themselves against an opponent. If they agree on something, they must find something else to disagree about or decide that agreement would be signaling something unfavorable, e.g. the argument that agreeing with an enemy on something signals acceptance of their side’s actions.

The question, going forward, is how should we think about groups and territory? You and I are able to identify individual sub-groups by who they disagree with, as well as the style and topics they tend to use and address. This, in itself, is interesting: how many groups can we name and recognize, and how many groups do we expect a single person to belong to online? Do identities have to show allegiance to one group primarily? Is that one reason for the prevalence of “alt” accounts?

The question of territory is much more dicey. Territory in virtual space seems to be largely determined by human attention, a concept that has been reoccurring in our conversations lately. We cannot directly measure human attention, though Twitter might be able to. Still, we have some sense that we know what people are thinking about. Are we fooling ourselves like the pollsters that are currently pulling their hair out, or is there something there?

Yours,

Wrench