Jacob Kaplan-Moss

Argument Clinic: Behind the Scenes

Intro: TKTK We wrote a play, the 7 takeaways [link], here’s some more stuff

Constraints and structure of the piece

This is the same argument, 3 times. Here’s how Jacob described it:

  • Thesis: the working relationship is more important than resolving this particular conflict.

  • Major story arc: two people have a niche and initial disagreement that gets out of control. They re-orient, and then leave understanding each others values agreeing that their relationship is more important than a disagreement and agreeing to support each other's values to the maximum ability possible they perhaps begin to resolve the technical conflict as well .

    • Scene one: disagreement. Disagreement escalates, until one comes close to crossing a line. Then they pause.
    • Scene two: initial, awkward, de-escalation, followed by a mutual discovery, that their working relationship is more important than whatever is going on here. They repair that working relationship, end with "we're good." (“Shakespeare nerd. I had been picturing the pyramid with the middle of scene two being the moment of the highest emotional tension.”)
    • Scene three: each re-states their values this time without argument, and with more mutual agreement, they identify the places where their values are aligned and where they're not they leave with a possible way forward

It’s the same starting point each time, but with different valences and outcomes.

  • But also it’s a bit jagged, for realism

  • Sumana notes: resolving the emotional conflict is maybe more important than resolving every detail of the intellectual conflict

No crossing the big lines. We deliberately didn’t want to demonstrate any behavior that would be a “call HR” moment in real life. The closest we got was raising our voices.

Fundamentally, this play was meant to be watched in the context of a professional conference (PyCon), so we wanted to avoid unpleasantly surprising or even triggering PyCon attendees. Our audience would show up to a session with the reasonable expectation of only witnessing professional conduct that does not cross any big bold lines. No one showing up to a talk at PyCon should have to watch people being abusive to each other.

One of the underlying lessons of this script is to resolve disagreements directly with the person you disagree with – i.e., escalation is a last resort – but that’s only true to a point. There’s a point at which someone’s behavior is so inappropriate that it’s no longer safe or wise to work with them directly. We didn’t want to show that, but it happens. Don’t take away from “Argument Clinic” that “you should never escalate”; instead, take away that what we saw here was perhaps the outer limit of what’s ok at work. Anything worse, and going to your manager, HR, or even an employment lawyer would be warranted.

As we said in one of our working discussions: “We needed to develop a set of conversational terms where the heat goes up and up, and by the end, we have come right up against the guardrail but not past it. These two people clearly are having a huge problem with each other and have not dealt with it productively. But there is a path back.” Which means: no unforgivable offenses. In a preview session, some of the audience found that Parvati’s sarcasm and condescension towards Manny in the first several minutes made it unbelievable that she was a beloved manager; we addressed this by Sumana reducing the acidity in how she delivered those lines.

What about gender and race? In real American tech companies, a work conflict between a white, Jewish man and an Indian-American woman probably feels fraught because racism, sexism, and anti-Semitism might play a part – and because you can never be sure that they aren’t. A brown woman would likely be concerned about being perceived as shrill, too angry, or too negative, and a Jewish person might worry that sharing concerns about the company’s finances would invoke stereotypes of greed. That is pretty heavy stuff, and, like abuse, not something we wanted to foist on unsuspecting conference audience members. Sumana thought of this as a bit like the world of Rosemary Kirstein’s Steerswoman novels: a bit of a fantasy world where people of all genders and races intermingle without their genders or races getting in the way of what people think they can do or be.

For accessibility and sensitivity, we limited how often Manny and Parvati literally speak at the same time (there’s only one full sentence of overlap), and gave the script to the CART captioners ahead of time to make it easier for them to live-caption the fast-moving, interruption-laden dialogue. We also changed one phrasing (“blindsided” to “ambushed”) to make it less ableist.

And, to keep things industry-conference-friendly, we removed profanity that went beyond mutters of “damn” or “hell” or “Oh my God”. This changed a few lines in particular:

  • Manny, upon hearing there’s no coffee: “Well, this is going to be a shitshow.”
  • calling a noisy channel “full of BS”
  • saying that some work “sucks” changed to “it’s a grind/a slog”
  • Manny’s “I didn’t think we’d be talking about my crap, either” changed to “I didn’t think I’d be admitting to you that my team’s unhappy with me”
  • “You did it without being a hardass” changed to “without being a dictator”
  • Parvati’s wry line (drawn from real life) was “that’s the most immigrant-ass thing I’ve ever heard”; we redacted the “ass”

Avoiding nerd-snipe-y details. We wanted to set up a realistic scenario so our audience could relate to it and put themselves in Parvati’s and Manny’s shoes. But we didn’t want to distract them with details about the company’s product, the Tollbooth library, what open source license they’ll use, and so on. In real life, probably our characters would share more of those details as they discussed the situation and made plans, but the audience would likely start thinking “is that the right solution?” or “what about [this factor]?” instead of staying focused on the characters’ behavior. (Similarly, in a previous tech conference play about code review approaches, Sumana realized that the audience does not actually need to see the code being reviewed.)

Length: 30 minutes. There’s only so much we could demonstrate. In real life this conflict would probably take longer than 30 minutes to resolve.

Also, we concentrated on one conflict from start to finish, so we couldn’t illustrate multiple kinds of conflicts. Many real-life conflicts aren’t between peers; they’re between people who have different amounts of power, including managers and their direct reports. Those conflicts often require different strategies.

Remote prep. Jacob, Kyle, and Sumana live in different places. We wrote, rehearsed, and revised this play over many hours of calls on Whereby and Google Meet, collaborating in Google Docs and Google Slides. Sometimes we improvised from an outline and captured an automated transcript using Fireflies.ai, then copied bits we liked into the draft script. At one point, to nail down what emotions and ideas characters would bring up and in which order, we used a Jamboard to arrange them as sticky notes.

Simplifying production. We didn’t have the full theatrical toolset available to us, so we had to try to write and act dialogue that would carry the whole thing, immersing the audience in the reality of the play. We had a fairly bare set, handheld mics, and only 1 backstage staffer (our director, who also ran our slides). No extra cast members to serve as narrator or secondary characters, no controls to dim the lights before and after the play or between scenes, and no big whiteboard or table. If we’d had more backstage staff, we might have added some sound: light background music for a few moments, or a few Slack pings or meeting notifications.

Friendship. Jacob and Sumana genuinely like each other and don’t like being mean to each other, and that made it harder to develop and rehearse the parts where Manny and Parvati do just that.

  • SECRET LINK to a SEPARATE SECRET WEBPAGE

  • This is, structurally, a romance. As we were developing the script, Sumana offhandedly said that we didn’t have to resolve the entire conflict between the two characters – they could end, as genre romance readers say, Happy For Now rather than Happily Ever After. As she said that she realized “oh crap, is this play a romance??” and Jacob wholeheartedly declared that it was.

  • Sumana remembers some articulations that romance author Courtney Milan wrote (but Twitter search is terrible so she can’t find them now): in the romance genre, characters earn their happy endings by bravely letting themselves be vulnerable; romance is a genre of justice in this way. The reader has to initially genuinely believe that these two characters CANNOT come together, because of irreconcilable differences, and then believe the changes that make their romance inevitable. So the author has to be like the God who makes a rock so heavy he can’t lift it. And then has to lift it.

  • hurt/comfort: fanfic trope: a way that you both show and have increased intimacy between two people is that one of them gets hurt and the other comforts them.

  • Manny feels like Jay is cheating on him, and “well he would have come to me if he really loved me… what is he getting from you that he isn’t getting at home?” [a lot of funny stuff about this in our discussion in Improv ¾ like around 38:05 and 43:52]

  • “That danger of, oh, boy, this feels, like, uncomfortable. It feels dangerous. Also have I overextended and now the other person is not going to reciprocate. Like, that's romance, baby.”

  • Manny is a bottom

Thank-yous

Thank you to

  • Kyle R. Conway, director, dramaturg, and audiovisual tech
  • Zed Lopez and Brendan Jercich for rehearsal help
  • Our online rehearsal audience: Andromeda Yelton, Elizabeth Yalkut, Erty Seidohl, Heidi Waterhouse, Candra K. Gill, and fhocutt
  • Our in-person rehearsal audience in Salt Lake City [TKTK]

Some lessons we couldn’t fit in

  • “Disagree, then commit”: a popular principle that says, as a manager, you can speak up about your disagreement with a potential decision, but once the decision’s been made, you should commit to implementing it and back it up, including supporting it in front of your team. This helps teams have a bias toward action, and socially, it means you don’t ruin your working relationship with your peers by undermining their decisions in front of colleagues.

A few weeks after “Argument Clinic”, the draft open source strategy that Manny and Parvati have written will likely cohere into a compromise that they can both live with. When they go back to their teams, “disagree, then commit” means they shouldn’t openly dwell on things they lost in the negotiations, but instead publicly support the compromise draft and work towards its success.

  • The tank: being a knight/tank, escalating professionally to a deliberate confrontation when your previous heads-ups/warnings/etc. have not worked to get someone to take a concern seriously

    • (“rental car dad”)
  • Thesis: the working relationship is more important than resolving this particular conflict.

  • https://github.com/jacobian/.org/issues/262

Deleted scenes

Some ideas we thought about, but didn’t go with:

“The meeting in 30 minutes.” We considered starting off with a more direct lead-in to the conflict, such as discussing the agenda for a regular management meeting that’s happening in 30 minutes, and one of them mentioning that resourcing Project Foo should be on there. They’d disagree, and realize they have to get on the same page, urgently, or else they’ll be messily arguing in the meeting (which is itself the last scene of the play). Or: at the start of the play, Manny’s 1:1 with Jay is in 30 minutes, so Manny and Parvati need to resolve their conflict before then.

Manny’s core insecurity is about his technical skill. That’s one reason he hasn’t been paying attention to that Slack channel. Parvati picks up on this and offers to help him with refreshers on the topics he doesn’t know well.

Starting Scene 2 with Manny and Parvati apologizing to each other, he for missing the prior discussion, she for having accused him of not caring about their shared values.

Open sourcing this tool might get amazoned/problems with investment

Puja asking, hey, your team is really fully booked, and I just want to make sure you explicitly know if there's any of that where you've been, like, waiting. On stuff from my team or there's some bit of it where you kind of secretly think this should be a platform like Task, please let me know

The Slack channel. Parvati goes into more detail on the problems with the Infrastructure Working Group’s Slack channel: it isn’t working as intended because some high-volume posters aren’t following thread etiquette. She doesn’t want this channel to go the way so many communication channels go, where what’s supposed to be a digest gets flooded, so many people stop paying attention, so yet another channel gets created, in a downward cycle. Relatedly, she and Manny negotiate about how he uses and will use it, mentioning that, since it’s full of technical details, he delegates his team leads to keep up with it. He’ll try to start reading it but he’d still appreciate a heads-up about things that he should pay particular attention to, in case he would otherwise miss them.

A more detailed discussion of what specific tasks need to happen as part of the project launch, what could be delayed till a future quarter, and how to rearrange the teams’ deliverables along the way

Passive aggressive exchange of “I don’t want to tell you how to manage your team, but….” Similarly:

Parvati: The fact that you missed this fairly important management level decision does not speak well of your ability to actually handle your responsibility.

Manny: The fact that you're doing management decisions in a channel labeled “infrastructure” doesn't speak well to your management abilities.

Manny literally came to this company partly because of Parvati; when he was looking for a new role and realized Parvati was here, he came here to learn from her, but “I felt like I needed to kind of get my own ducks in a row first before I could ask you for anything.”

Parvati offering to literally have the alignment conversation with Jay on Manny’s behalf, or to sit in on the conversation.

A joke about Manny’s love of process:

Parvati: The conversation that we're having right now, and the messiness, and the, like, undercommunication from having an argument at the coffee pot, it makes me think that this is probably worth treating not just as a one-off, but as if I really want the organization to commit to not just throwing this over the wall, but to contributing to open source. Because if it is and has been a part of our company values, then that needs to be owned as a responsibility by somebody. Right?

Manny: It easily could still be you, but you could have more resources or a defined amount of resources for it. Like there could be a, a more centralized decision making process about that. I know that's not as fun as, like, the early start up kind of “do whatever you want” ethos, but that may be a value that you might have to sacrifice to stick with the open source value.

Parvati: Manny, are you talking about how process actually makes it easier in the long run? Again?

Manny: Of course I am.

On Parvati disclosing vulnerability and needing support:

Jacob: All right. Okay. I mean, whatever. We'll give it a try. Worst thing that happens is I get the deer in the headlights look for a while, figure it out.

Sumana: You would not be the first man in history who has been asked for support from a woman and gotten “deer in the headlights.”

Jacob: I'm sure I have no idea what you're talking about, and you are not allowed to ask any of my partners that question.

Ending the play with Manny and Parvati still completely disagreeing about whether to release Project Tollbooth, but doing so more cordially, their working relationship resolved

Ending the play with Parvati saying, since she’s going to grab some coffee from a coffeeshop, she could get Manny a cup too, in exchange for his help with the JIRA tickets: “Manny, I am battling between my desire to make fun of you endlessly for loving JIRA and my desire to buy you coffee so that you will make me those tickets.”

Both characters break the fourth wall to explain their takeaways directly to the audience

Parvati’s name used to be Pooja, and Tollbooth used to be “Ringfire” or “Panda Zodiac”

The script itself

[export a final version of the script to PDF and stick it someplace downloadable, CC-BY-SA]