“Hey Jack!”

I’m Jack. 

It’s 5:11 PM on a Friday as I walk into Paesano’s, with the door closing behind me. The snow just outside the door is dry and crunchy. Stu is standing a few feet away where the entry hallway meets the main hallway, smiling at me. He’s wearing a puffer jacket with a dusting of snow on the shoulders, and his cheeks are red, probably from the cold. He must have just walked in himself – say, 5:11 PM like me, or 5:10 PM at the earliest.

Stu is a colleague who I am surprised to see here. I’m here every night, because Paesano’s has a good Chicken Marsala, so every night I have the Chicken Marsala before I go home to my apartment. That is optimal. The optimal time to come on Fridays is 5:10 pm to 5:15 pm, and it is 5:11pm as I am walking in. On Fridays the restaurant turns into a happy-hour spot and then sort of a nightspot, but I’m always home before any of that.

Stu is a grad student like me. His advisor is more famous than mine and has better funding. I should stay in touch with Stu in case I need to switch advisors. I’m surprised to see Stu because I live on the outskirts where University shades into Town, and people from my department don’t come here often. Stu’s presence needs explanation.

I know that Stu will expect a response. What are my options? Non-exhaustively:

1) don’t say anything, 2) pretend I didn’t see him, 3) feign unconsciousness, 4) act like I didn’t recognize him outside the university context, 5) pretend to be ill, like I am having a seizure, 6) walk in the opposite direction, 7) rebuff him directly (“Don’t talk to me!”), 8) catch him up on my most recent experimental results, 9) ask him if his advisor is looking for students to join his lab, 10) craft some kind of social small-talk-like response and put it back in his court and see what happens next. 

I should do any of these with some nonzero likelihood, with some being a lot more likely than others. My estimates: 1) 0.03%, 2) 0.05%, 3) 0.09%, 4) 3.03%, 5) 3.9%. 6) 5%, 7) 7%, 8) 12%, 9) 33%, 10) 36%. Frustratingly I am caught without a good source of randomness. There is no visible clock. The colors of the wallpaper, curtains, carpet, and servers’ uniforms are the same as they are every night. I could pull out my phone and randomize on the seconds digits of the clock time, but that would take long enough that it would be its own social action: 11) pull out phone and look at it before responding. I decide to just go with the dominant action this time, while vowing not to make it a habit. Your adversaries can exploit you if you become too predictable.

So it’s social small-talk-like response time. What are my goals? #1 is to fit in and not seem peculiar or eccentric. I know that the way I think is atypical, but best to minimize the obviousness of that. #2 is to improve relations with Stu, in case I have to switch advisors and need an in to his circle and his advisor. #3 – maybe I can learn why Stu is here. 

The easiest way to not seem eccentric in small-talk-land is to mimic. I could just come right back with “Hey Stu!”, in parallel with Stu’s “Hey Jack!”, but I’ve made enough progress recently with small talk that I know that it’s more advanced to switch things up and not repeat the other person verbatim. If the other person says “Hi”, then “Hello” might be slightly preferred as a response, and vice versa. Also I understand that a  differentiated response can move the conversation along (though I don’t always know what goal we are supposed to be getting closer to by moving a small talk conversation along).

Here’s an idea: differentiate my response by combining it with my desire for information about why Stu is here. Something like “Stu – what are you doing here?” Add some slight intonation of surprise and pleasure, along with a medium welcoming smile – not too try-hard, but still communicate that Stu’s presence is a positive, to keep up the rapport. I might need to switch advisors.

Now, on to intonation – which words should I emphasize? I will emphasize “Stu!” with the intonation of surprise and pleasure. For simplicity let’s assume that the emphasis of the other five words are binary yes-or-no for each word, so 2^5= 32 possibilities. Some possibilities can be discarded immediately – for example, emphasizing all five words is the same as speaking loudly without emphasis, and emphasizing no words is just speaking softly without emphasis. That brings it down to 30 possibilities.

Emphasis of each word changes the meaning. I am not sure what effect emphasizing “What” has. Emphasizing “are” sounds English-upper-class and privileged. Emphasizing “you” might seem like I am questioning Stu’s rights to be at this restaurant, as though he is uniquely unqualified to be there, unlike me and all the other guests. Emphasizing “doing” is headed in the right direction of social curiosity, but it also carries with it some obligation on Stu to account not just for his presence, but for his activities. Emphasizing “here” fits well with the facts and the informational goal – I wouldn’t say this sentence at all if I had just run into Stu at the lab, because that happens frequently. It is really the location that is anomalous and there’s nothing else that I want to inquire about.

“Stu! – what are you doing here?”

I gave that reply without missing a beat. I am pleased. I think I hit that one out of the park. Lots of considerations and calculations to balance, but I completed them all within the time bounds of social back-and-forth. Probably about 1.5 seconds.

As I said, I know that my thought processes are atypical, especially with regard to social interaction, but I am an extremely fast thinker. I think that anyone in my department would tell you that. I can do complicated arithmetic calculations in the time it takes most people to return a ping-pong serve. I can read probability landscapes and payoff matrices off the back of my eyelids.

Stu: “Oh, you know, end of a long week – everyone was feeling like it would be good to get out, get off campus for once. A bunch of folks should be coming, we were going to get some dinner. Do you know this place? I’ve never been here before.”

Lots to unpack here, all at once. Who is “everyone”? And is “everyone” the same as “a bunch of folks”? Probably “a bunch of folks” is a subset of “everyone”, because you would need to be known to be interested in getting off campus before you committed to a group effort of actually getting off campus. But “everyone” is the interesting question. There are three experimental labs in the department – there is some socializing between the labs, but also folks in each lab primarily talk to each other. I’m guessing that Stu’s “everyone” means all or most of the people in his lab, but not in the other labs. And (most important of all) would everyone include Stu’s advisor Louis? If I could talk to Louis, that could be important, especially if I end up needing to switch advisors.

But to Stu’s question to me – do I know this place? This is the kind of juncture that leads to uncontrollable branching of conversational threads, and irreducible chess-like  complexity. As I said, I am a fast thinker, but thinking about optimizing over five or six ply of back-and-forth lookahead with unknown probability distributions from my conversational partner starts to make me anxious, especially because there are also many different goals that I might try to optimize for. I can visualize the branches – they are black lines against a vivid yellow background – sparse at first, and then a dense tangle, merging off into the all-black distance.

Back to my goals: a) fit in (conceal my atypicality), b) build a rapport with Stu (because I might need to switch advisors), c) maximize my chances of meeting Louis (because I might need to switch advisors). Also a meta-level goal – I have to control the conversational branching factor or my anxiety may become uncontrollable and lead to me behaving in a clearly atypical way. Things I could say:

1. “I come here every night and have the Chicken Marsala and then I go home to my apartment.” (The most accurate, but high atypicality and lots of offered information with many branches from there)
2. “Oh yeah, a couple of times.” (Vague, low branching factor from here)
3. “Oh yeah, I’ve been here a few times – you know I live in the neighborhood” (a new set of branches from here. Do I like that set? I’m not sure how to evaluate it).
4. “Me too – never been here before, but I’ve heard good things about the Chicken Marsala”. (Good solidarity, fitting in, and introducing a new topic, but one where I feel comfortable in my expertise.)
5. “Same here – my first time”. (A high degree of fitting in with Stu, low accuracy, low atypicality-reveal, low branching factor.)

Me: “Same here – my first time.”

This is OK, but I don’t think that it’s great. It’s not like my first reply, which was clearly stellar. There were too many goals pushing in exactly opposite directions this time – rapport-building pushes in one direction, and fitting-in / low branching factor / low atypicality-reveal push in the other. I know that this one-dimensional opposition is simplistic, and that there’s probably a sweet spot, a peak in this high-dimensional landscape that simultaneously reduces branching, increases rapport, decreases atypicality-reveal, and optimizes the chances of talking to Louis. But my searches out three-or-four-ply of high-likelihood exchanges weren’t finding it. 

The background of the branching diagram is like a vivid neon yellow. I know I’m not literally looking at it, but it still hurts my eyes a little bit. I have a faint foreshadowing of a headache, but at least it doesn’t feel like it will be a migraine. My migraines can be really bad.

There’s something that bothers me about my response. I don’t see any red flags or checkmates in the near-term back-and-forth response tree, but I have this feeling that there is some danger there that is seven moves out or so. It’s like when you make an easy piece capture in chess, but have to wonder if it was a sacrifice – the beginning of a gambit. I experience this anxiety as a clicking noise – sparse, occasional clicks, like a Geiger counter.

Stu: “Cool, cool. Well, I’m early – do you want to check out the bar? I’ve got a few minutes before my people start to get here.”

This is a completely different response space, with a huge number of vectors and choices and possible wordings, and none are anything I had gamed out in advance, because I didn’t see this coming. 

Possible responses stream past me randomly, like confetti in front of a high-speed fan pointed at my face. Some are improvisations, some are lines that come to mind from movies and sitcoms and Westerns, and long-arc television dramas, and hard-boiled private-eye stories from the 1930’s, 1940’s, 1950’s, and 1960’s:

“I don’t drink”, “They have a bar?”, “Sorry [holding up phone], have to take this”, “It’s 5 o’clock somewhere!”, “Yeah man totally”, “Not today Stu – doctor’s orders!”, “Sounds like a plan!”, “We can check but I’m sure the bar’s fine!”, “Let’s do it”, “Shaken or stirred haha?”,  “I’m really more comfortable here in the hallway”, “I think my table’s ready in the dining room”, “Wouldn’t want to keep the ladies waitin’!”, “Oh, I was just on my way out actually”.

I spend about 0.4 seconds just letting the confetti blow past my head, trying to think what kind of directionality or principal vectors or subspaces I could find to put some useful structure on it. Then I realize that the responses are probably easily clustered. I try clustering a few just by similarity of intent and result, and it starts getting clearer. Yes, there are a few that don’t quite fit any clusters (jokes, and questions, and completely unintelligible grunting and muttering), but almost every response fits into one of two clusters: “Yes”, and “No”. Some responses mean that I am willing to go into the bar, and others mean that I am not willing to go into the bar. It may be that that is the whole point, and the form of Stu’s question doesn’t matter much either. He is just trying to elicit my announcement of agreement or disagreement. I could as well just say “1” meaning yes, or “0” meaning no. In fact, Stu is technical enough that I bet if I said “1” he would be initially puzzled, but then take it as a geek joke where 1 obviously must mean yes where 0 means no, and then we would go into the bar.

So: first, choose a cluster, then secondly choose a wording from the cluster. Which cluster? They separate entirely as they branch to the future. The end states on the paths into the Yes cluster and the No cluster are almost entirely disjoint. There is a clear path of bright yellow in between them, but that’s not a path I can walk on – it’s yellow because there are no paths. The entrance to the bar is like an event horizon – sure, I will physically come out again, but I can’t undo having gone in.

Focus, calm, optimize. Think big-picture. Branching factor is past the event horizon, can’t estimate it. Fit in, minimize atypicality, maximize chance of talking to Louis. 

Me: “Sounds like a plan!”

I’ve never been in the bar, just the dining area. The bar is parallel to the hallway, with a hallway entrance at one end and a dead end at the other, with a large clear window fronting onto the hallway. Just a rectangular room big enough to hold a bar and room enough behind the black fake-leather barstools for patrons to easily move around. Stu leads the way. No one is sitting at the bar. There are 11 stools – 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, in the order that you encounter them from the hallway. There is a clock or cash register or something at the far end of the bar that is blinking faint blue digits about once a second, which is distracting. I will try not to focus on it.

Stu walks in ahead of me and sits in stool #6 – the exact median stool. Nicely done. Now, what stool is best for me? Do men’s room urinal rules apply here? If so, I should choose #1 or #11. But this is an interestingly different situation. Stu and I will probably be talking as opposed to not-talking as in men’s rooms. I should choose a nearby stool. #5 and #7 seem like overdoing it in the closeness department. I choose #4, which seems slightly too far away, but then I split the difference by sitting on the edge of #4 that’s closest to #6, and kind of leaning toward Stu in addition. 

The bartender asks Stu what he’ll have, Stu names a cocktail I’ve never heard of. I hold up two fingers, which I understand means “I’ll have what he’s having”. (“Fit in”).

Stu: “Good to see you again Jack. I don’t think we’ve really talked since the September ice-breaker thing.”

This surprises me. Stu is correct that we had an extended conversation at the cross-laboratory intra-departmental ice-breaker social meetup event for graduate students and faculty on Friday September 6th in the tea room adjacent to the Lederman lecture room. But more recently than that we talked at the intra-departmental external colloquium lecture (informally known by the grad students as the Job Talk Talk, as it’s usually delivered by university-external speakers who aspire to tenure-track appointments within the department) on Thursday October 10th in the Baker Auditorium in the northwest corner of Hinton Hall. I had asked a question near the end of the Q&A period of the talk that demonstrated that the speaker was wrong, and moreover that he was wrong in more than a trivial and ignorable way. Stu followed up with me and asked me about the assumptions behind my question afterward, which (although I disagreed with his premises) I thought was a very reasonable question to ask me.

So how can Stu think that we have not really talked since the September ice-breaker event? Alternatives: 

1. Stu is calendrically confused. He thinks that the ice-breaker conversation happened after the colloquium conversation, so the ice-breaker conversation is the most recent in his mind.
2. Stu is implicitly applying a filter on our conversations represented by the “really” in “really talked”, with the set of conversations that survive that filter being those that are sorted by time to find the most recent. Our September ice-breaker conversation makes the cut where our October colloquium interchange does not. It’s true that the September conversation was longer (15 minutes?) where the October conversation was shorter (3 minutes?). Also the September conversation included some generalized back-and-forth about personal details and backgrounds and aspirations, where the October conversation was more narrowly focused on technical issues. To my mind, the “real” in “really talked” is a better match to the shorter technical conversation than to the longer social conversation, but I don’t know what Stu’s version of the “really talked” filter would look like.
3. Stu simply does not remember our interchange at the Job Talk Talk colloquium. This seems unlikely because I think the exchange was memorable.

I tentatively conclude that #1 is the most likely explanation – Stu is incorrect about the calendrical ordering. This observation is worth filing away in case we ever collaborate.

[Thump!] 

The bartender puts down a glass of ice water plus a coaster in front of me, and then does exactly the same in front of Stu.

I have to admire the efficiency of this. The bartender does not set out a coaster for me and then place a glass of ice water on it, and then set out a coaster for Stu and then place a glass of ice water on it. And he doesn’t set out two coasters – one for me and one for Stu – and then place ice-water glasses on top of each of the coasters in sequence. No – with one hand (the right one) he puts a coaster plus ice-water unit in front of me, and then again with the right hand a coaster plus ice-water unit in front of Stu. I don’t know exactly how he is delivering the individual units, but I suspect that he somehow fills an ice-water glass, which may require both of his hands, and then starts the motion of delivering that glass to its destination in front of the customer, but along that path somehow scoops up a coaster from some stack of coasters, and folds the coaster under the glass, and then delivers the whole thing as a package, all with the same hand.

This seems like the kind of habit that has been optimized for being fast. I’ll bet that some of the time it’s really important for bartenders to be fast, but right now, with Stu and me as his only customers, he could slow down a little I think. He could pick up a coaster and drop it in front of me, and then kind of drawl “How we doing tonight, gentlemen?”, and then put a coaster in front of Stu, and then say something about the dry crunchy snow we’ve been accumulating over the last couple of days, and then move on to getting the glasses of ice-water together. But maybe it’s just a habit he developed through all the times when he really had to move fast because as they say the joint was jumpin’, and now it’s his habit and it’s just how he does it, without thinking. Actually, as someone who has never worked in a bar, I don’t even know what the demands on his time are beyond catering to his immediate guests such as Stu and myself, and what kinds of multitasking duties he has with regard to glassware cleanup and pre-mixing pre-mixed cocktails, and setting up the various containers of olives and cherries and garnishes for drinks that will come later in the happy-hour rush. Maybe every second that he saves in efficiently setting out the complimentary glasses of ice water is a second he can spend using that little grater tool to shave off spirals of peel from oranges and lemons to put on the rims of cocktail glasses.

The ice-water glass in front of me is sitting on a coaster made of some indeterminate kind of white plastic foam, and the coaster probably has a branding logo of some kind on the top, but if so I can’t see it because of the ice-water glass. The bar itself is some kind of white stone (marble?), with an enclosing metal trim (stainless steel?) running along the top front edge of the bar. That seems unusual – aren’t bar tops usually made of wood? And I know that marble stains easily, so not an ideal surface for serving red wine and brown liquor. No, it’s not marble, that looks like quartzite. Quartzite! Not my absolute all-time favorite stone in the whole world, but definitely my favorite metamorphic stone in the whole world. And it has to be really really old. The slab is probably 1.7 billion or 1.8 billion years old. Dizzyingly old, and the oldest thing in the bar by far. This chunk of quartzite sat there, buried for longer than a tenth the age of the entire universe until, just a few years ago, someone dug it out and cut it out into the shape of a bar and hauled it over here and installed it at Paesano’s.

The bartop is definitely the oldest thing in the bar, so max(age). What’s average(age)? That depends on how we segment everything out, and decide what counts as a thing, I guess, and then when those things started their life. It’s easy to feel while I’m talking to Stu that it’s just me, Stu, and the bar – so me (34 years old), Stu (28 years old?) and the bar (say, 1,800,000,000 years old?), for an average of 600,000,021 years (last few digits not significant). But there are a lot more things in the bar than that. There’s the bartender (45?), and 11 stools (which look a little beaten up – maybe 15 years old each?) That already brings us down to an average age of 120,000,018 (last digits not significant again).

There are so many things in the bar, though, and you don’t know if they should be counted individually or grouped and counted. The glasses, the coasters, the cocktail napkins, the plastic toothpicks, the individual tiles on the floor behind the bar. The back of the bar, with the rows of bottles of liquor stretching up to the ceiling, above where anyone could realistically reach without a ladder – is that one big thing or a collection of individual bottles? Helpfully some of the bottles of liquor say right on the bottle how old they are (12 years, 15 years, 18 years), but you have to be immediately suspicious of that. Are those numbers up to date? How long have the bottles been on the shelf? 

Really the only thing that matters is how many non-quartzite things there are, not how old the non-quartzite things are. The recent age of non-quartzite things just injects rounding error. If there are 1000 things in the bar, including the bar, then the average age of things in the bar is still 1,800,000 years old, even if everything other than the bar is zero years old.

If people would just agree on what qualifies as a thing, and how the things should be grouped and averaged, then we could put bar codes or QR codes on all the things, encoding the kind of thing it is and when it started out. Then you could automate the whole calculation. It could even be a smartphone app – scan the entire room, bam, done, you’ve got your average age. It could even update itself, one second per second, without rescanning and recalculating anything, until some particular thing like a really old bottle of Scotch went away, or a shift changed and the 45-year-old bartender was replaced by a 26-year-old bartender. 

I hope that someone in some academic department is working on the problem of what counts as an individual thing, because that’s the first part of the problem. But I’m a realist – we’re not going to get everyone to agree on what the things should be very soon, and even then setting up the bar codes and associated software infrastructure will take a while. The average age of things in the bar is a hard problem. It may be harder than is really worth it to think about in detail, especially right now. 

[Wacka-wacka-wacka-wacka-wacka!]

I’m startled for the first half-second or so, but then I recognize the sound, and remember having heard it in the distance sometimes on weekend nights when I am having my Chicken Marsala in the dining area. The bartender seems to be putting in a lot of effort and muscle as he shakes the cocktail shaker. I wouldn’t think it would be that heavy or that difficult, but maybe it’s harder to do than I think. 

The bartender looks very serious as he does his cocktail-shaker-shaking task. You might think he’d be amused at the silly ritual, or at least project a festive air. But he is impassive and focused, with his eyes seemingly locked on the corner of the ceiling to the left behind us. His body is facing straight ahead to his side of the bar, but his head is stable at about 30° yaw and 25° pitch. Roll is negligible, which makes sense if you think about it. It’s not hard to roll your head from the default looking-straight-ahead position  (like some people from the subcontinent do expressively as a kind of nodding gesture which most Americans don’t have as part of their nodding signal set, which is limited to pitch for yes and yaw for no), but it’s not so easy to roll your head if you’ve already got some significant yaw and pitch going on. If you tried to get to, say, 30° yaw / 25° pitch / 20° roll, you could hurt your neck.

I have to remember that my atypicality gets revealed not only by unusual things that I say but also by atypical response times, because I sometimes become overly absorbed in trains of thought that are not directly connected to the social conversation. For the first interchange or two with Stu, my responses were snappy and within the expected time bounds, but now I am not even sure how long it has been since my last response – 4 seconds, 8 seconds, longer? 60 seconds, 120 seconds? Five minutes?? This is not good! If you’re playing ping-pong you’re required to respond quickly or lose the point, and even if you lose the point you’re expected to start another point soon.

Question: What kind of ping-pong player catches the ball in their hand in the middle of a point, and then puts the ball in their pocket and walks away?
Answer: an atypical ping-pong player.

Stu: “It’s funny how these things get momentum. We were just chatting about getting out, and a few people pile on, and before you know it Louis is jumping on too. I think he’s even arranged a room here for us for dinner.”

This kind of pronoun vagueness (“we”, “us”) irritates me. I know that legal documents sometimes have a Defined Terms section at the very beginning, so that it’s clear for the rest of the document what a term like “the Company” means. Personal conversations should have those too. This should be legally enforced. It should be a citizen’s right to say “Stop! You are using a term to refer to a group of people, but which people are covered by the term is not clear!” And if the person continues to use the term without clarifying then there should be sanctions and penalties, enforced by either the government or appropriate civil entities.

When I was working at the tech company before grad school there was a very senior computational linguist named Dmitri, who everyone was a little bit afraid of. One day Dmitri was running an engineering meeting, with about ten people in the conference room. Everyone in the room was an employee of company X which had recently  acquired company Y. Everyone in the room had also been an employee of acquired company Y, which is why they were there, and why they were now employees of company X. And as he led into some anecdote or other, Dmitri prefaced it by saying:

“Ever since we (making a sweeping hand gesture of inclusiveness toward all the people in the conference room) acquired … us (making the same gesture) ….”  

For some reason, this didn’t irritate me at the time. Maybe it’s because everyone knew what Dmitri meant, and there was no ambiguity. Or maybe it’s because Dmitri, as an accomplished and very senior computational linguist, would be very familiar with the knotty challenges of coreference resolution, and to the extent that he was using these pronouns in ambiguous and possibly-conflicting ways he was doing so with full awareness and intentional irony, as though to point out a paradox of reference as a little side display on the way to his broader point. Dmitri could be kind of a show-off. Also I didn’t have to worry about whether “we” and “us” included me. I knew that I was part of “we” and I was also part of “us”, even though “we” and “us” referred to different sets of people.

New voice: “Is this seat taken?”

I turn to the left and see that the new voice is Hannah, walking up to us. Hannah is also in Stu’s lab, and they are very close friends. People say that Hannah is Stu’s “lab wife” and that Stu is Hannah’s “lab husband”, even though Stu has a real-life wife who lives with Stu and doesn’t work in the lab, and Hannah has a real-life husband who lives with Hannah and doesn’t work in the lab either. 

Stu laughs, and makes a welcoming gesture with his right hand, and Hannah … sits down in seat #5! Right between me and Stu!

This is several different kinds of bad news. For one thing, how am I supposed to talk to Stu now, and build rapport with Stu so that I have a connection to Louis’s lab if I do need to switch advisors? I quickly think back over my conversation with Stu thus far. I have said a total of three things to Stu: “Stu! What are you doing here?”, “Same here – my first time”, and “Sounds like a plan!”, and all of them were before we even came into the bar. Man, I’ve really put the ping-pong ball in my pocket this time. If the next time I run into Stu is in the spring, he’ll probably say something like “Good to see you again Jack. I don’t think we’ve really talked since the September ice-breaker thing.”

Hannah herself is her own kind of bad news. The last time I ran into Hannah I was … not at my social best, and my conversational flow and my choice of topics were very obviously not typical. Since then I have preferred not to think much about the interaction we had. Now I am forced to think about it again. Also the geiger-counter clicking that started when I said “Same here” to Stu is suddenly clicking again and getting faster and louder.

Hannah is turned toward Stu and away from me, as she and Stu seemingly share some joke or other. In fact, for a moment her head almost perfectly occludes my view of Stu’s head, just like the Earth almost perfectly occludes the Sun in a lunar eclipse and the Moon almost perfectly occludes the Sun in a solar eclipse. I wouldn’t have predicted this. Hannah is fairly tall, but not quite as tall as either Stu or myself. I wouldn’t have expected that our heads would line up so perfectly. But probably most of the height difference is in our legs, and we’re all pretty much the same height when we’re sitting on barstools. Also my posture is not the greatest, and neither is Stu’s, and we’re all settled in and a bit hunched over the bar, and maybe that’s why there is this moment of perfect alignment, which I assume will be temporary.

Hannah: “Jack! Wow, you’re quite the Paesano’s regular, aren’t you? Seems like every time that I’m here you’re here too, having dinner!”

Hannah has turned toward me now, and is smiling, though I don’t feel like it is the smile of someone who genuinely wishes their conversational partner well. And then, like Earthrise on the Moon, Stu’s head slowly comes into view behind Hannah’s head as he leans forward toward the bar, and turns toward me, looking at me, his eyebrows raised. His forehead looks like … I know that there is some old-timey agricultural phrase that describes a forehead that looks like that, with lines like would be left by a plow? Furrows? That’s it – his brow is furrowed.

I start my mental conversational-response-generating machine, expecting to take just a moment to generate responses and do some clustering and assign probabilities and make a somewhat-randomized choice, and … I got nothing. Nothing is coming out. It’s as though responses are running into each other and conflicting and canceling each other out as some lower level of the architecture, and almost no candidate responses make their way to my conscious awareness. I briefly consider “Madam, I have never seen you before in my life!”, and reject it as implausible. I think of “I was just saying to Stu that this is my first time at Paesano’s … today! (haha)”, and that doesn’t feel right either.

I wish I could talk to Stu directly, but that’s not possible in the everyone-sitting-in-a-line-at-the-bar configuration. How do people communicate in this situation! Bars are supposed to be situations of high sociability, and long-term committed couples are always saying things like “How did we first meet? Oh, we met at the bar”. How does that happen? Are meetings that happen at a bar always between people who randomly ended up in seats next to each other?

It’s not like non-adjacent people cannot communicate with each other at all. There are three communication paths that I can think of. First, you could shout something that everyone could hear, if the bar isn’t too loud, and someone could shout back. I could shout “Go Blue!” and probably some non-adjacent person at the bar would shout “Go Blue!” in response, even if there isn’t any sporting event on the TV at the moment. This is low-bandwidth communication and completely insecure. Secondly, you can in fact distort the spatial alignment a little bit, and create pockets of 2D communication, where two people in seats N and N+2 lean forward somewhat, and the person in seat N+1 sits back – sort of how Stu and Hannah and I are arranged right now, as Stu and Hannah wait for me to say something in response to Hannah. This can work, but might eventually be physically uncomfortable, and also doesn’t offer a way for the person in seat N to securely communicate with the person in seat N+2 (as I wish I could do with Stu right now) without seat N+1 listening in. It also doesn’t really scale beyond 3 people, maybe 4 at the outside.

The third option is to embrace the 1-D constraint, and have all direct communication happening between adjacent seats only. If someone wrote a science fiction book about this social world, it wouldn’t be called Flatland – it would be Lineland. Non-adjacent people can still communicate, but it has to be by message-passing. Seat #2 says to seat #3: “Can you let seat #8 know that I really like her hat?”. Seat #3 says to seat #4: “Seat #2 says that he likes seat #8’s hat. Eventually seat #7 says to seat #8: “Seat #2 likes your hat”, and seat #8 says to seat #7: “Tell seat #2 that I have a boyfriend”, and seat #7 tells seat #6 to forward that message on down. Pretty slow. And can more than one message be passed at once, or is that too much? Can they cross, or is that too much to remember? Can seat #6 tell seat #5 to pass on I-have-a-boyfriend from #8 to #2, to which seat #5 replies ‘Sure, I’ll pass that along, but also: “Go Blue!” to seat #7 from seat #1’. And this is all informationally insecure. Seats 3,4,5,6,7 now all know that seat #8 has a boyfriend, which might seem that it reveals only information about seat #8, but it also reveals information about seat #8’s reaction to seat #2 specifically (would the response be the same to seat #3 or seat #1?), and seat #2 might actually understandably feel that this reveals hurtful information about seat #2 himself, and feel that he should not have expressed his sartorial or millinerian opinion in the first place.

And all that is without alcohol. What happens to message-passing once patrons have begun to enjoy their Chardonnays and their IPAs? What happens when the word “hat” becomes “rat”, “flat” or “fat”? Maybe just confusion and some hurt feelings, but maybe this is how bar fights start. 

I look up and down the bar line thinking about message-passing, maximum transmission  capacities, robustness, failure modes, resend strategies, information security. I am also still working on an appropriate message for #5 and #6, but I have nothing good enough to send to them yet. I notice that the bar has begun to fill in. Seats #7 and #8 are taken by two guys that I can’t put names to but are definitely first-years in Louis’s lab. In the other direction are Jeremy and Asha, seats #2 and #1, respectively, also Louis students. And down at the end, just walking in the doorway of the bar, is Louis! The man himself!!

Louis: “Hey party people!”

Louis has a cocktail in his hand as he is walking in, which looks exactly like the cocktails that are in front of me and Stu – stemmed glass, brown liquid, a toothpick or something pinning the cherry to the bottom of the glass. Now how does he possibly have a cocktail already?! He is just walking in! Stu and I were the first ones in the bar, right? I don’t get it. Does Paesano’s have some secret university-faculty-only bar where professors can show up at 3pm and then only later on join their grad-student parties in progress at the regular bar? Or did he get here before any of us, and order from the regular bar and then retire temporarily to … where, that secret university-faculty-only lounge? Or did he call ahead and say “My usual please”, which cocktail Paesano’s wouldn’t have to ask further about and would have waiting for him at the door even though the last time he visited was a year and a half ago? But anyway that’s Louis all the way, that’s how he rolls.

As he walks in, I’m sort of on pins and needles to see how he deals with seat choice. What’s it going to be Louis? The available seats are 3, 9, 10, 11. He can choose seat #3 or seat #9 if he wants to be in maximal contact with (one of) his students. Or he can post up in seat #11, which arguably could be taken to be the head of the table, maybe put his feet up on seat #10, and then (with some definite distance between him and the rest of us) clear his throat and raise his voice a bit, and say something significant and leadership-like.

Instead, Louis walks all the way up to somewhere between Stu and Hannah, sort of plants his butt against the back wall opposite the stools, but remains standing, and waves super-cheerfully to everyone at the bar. OMG. Do you see what he did there? He had tickets to four seats in Lineland, and he declined the ride. No message-passing for Louis – he’s in full one-to-many broadcast mode. He can see everyone and talk to everyone in Lineland and everyone in Lineland can see him and talk to him. That is how you do it!

Louis: “Hey Hannah, Stu. A couple of updates – mostly good news. For one thing, the Oslo panel is on! And Hannah, I was thinking that you’d be perfect to represent us there, if you’re up for it. Also, I got the reviewer’s reports back for the STCO article, and just a few things to work through, nothing major.” 

Louis’s gaze drifts to me. 

Louis: ”Jack, right? You’re one of Vivenne’s?”

Louis: “Nice to see you. Looks like you fell in with a bad crowd tonight, with the likes of young Stuart here. How did that happen? I bet you’re questioning your life choices now.”

Louis seems to be saying something negative about Stu, but Louis is smiling, and Stu then matches his smile. They seem very friendly with each other, which is confusing. Confetti responses are blowing furiously past my head again, but as they do I manage to nod in agreement with Louis. Then as soon as I start nodding I realize there’s ambiguity in what I am nodding at. I meant to nod affirmatively that I was indeed one of Vivienne’s (Vivienne is my advisor), but soon it began to appear that I was nodding in agreement that I was questioning my life choices now due to having fallen in with the likes of Stu, which is not at all what I meant! I couldn’t switch to shaking my head negatively, because that might look like I was trying to indicate that Vivienne was not actually my advisor. Anyway, I stopped nodding immediately at that point, better safe than sorry.

Louis: “Coincidentally, Jack, your name came up just today in one of our review meetings. Vivienne was filling us in on your progress.”

Oh. Wow. This is important. I need to figure out what Louis really means here, and I have to concentrate to extract all the information I can.

What does Louis’s intonation tell me beyond what he is saying on the surface? I replay the sounds of his last two sentences in my head, and I listen for relative emphasis between the words, but I can’t really hear any particular emphasis at all. Each word seems to be at pretty much the same volume as the last, all sort of flat in tone, and they’re all spaced pretty evenly and normally, no hesitations. I also looked at Louis’s facial expression as he was talking, and I got nothing. He’s not smiling, not frowning, not laughing, not angry, not critical, not negative, not positive – just sort of impassive, like a poker player announcing a bet.

In some sense this has all gone unexpectedly well, given the goals I had coming in. When will I ever get a better opportunity than this?

Me: “That’s great, Louis. I’m glad that Vivienne was filling you in, yes, the progress has been excellent, more than excellent, actually …

I paused for a moment, as what came next could not be simple or brief. It was no longer a matter of choosing between canned phrases – this would be me taking the conversational floor for an extended period. I would have to choose the first sentences even before I knew what the last sentences would end up being.

The key points I needed to make:

1. That this important research problem has been long-neglected and yet there is so much opportunity
2. That I am making great progress on this challenging work and (by extension) I am a promising person worthy of investment
3. That it is sort of an accident that I am in Vivienne’s lab, in that there is no logical connection between my work and the other work in her lab. This is delicate! I cannot be seen to be disrespectful of Vivienne! I plan to address this by interleaving: I will make the case for my work not needing to be in Vivienne’s lab, then say nice things about Vivienne for a while, then continue the case, then say more nice things about Vivienne, and repeat that as many times as necessary.
4. Express my enthusiasm for Louis and his lab and inquire about transfer possibilities.

A final delicacy: for Louis to appreciate my progress on the main problem he needs to understand the issues in some depth, including traditional methods, baseline efficiencies, recent research results, inconsistency of funding, and so on. Although everyone knows the basics of ethylene oxide catalysis, it’s possible that he is not entirely up to date, and might need a refresher. I must feel my way – I have to establish common knowledge, while being careful never to lecture Louis, especially on material that he may already know. And I must also be brief! Concision is key! I must spend no more than five minutes on the refresher portion of my pitch.

And so I began to explain.

“… and it would be great to have a chance to work with Stu and Hannah and Jeremy and Asha and the two guys there in seats #7 and #8, whose names escape me right now, but who I’ve definitely seen around. What do you think?”

I stop talking, and no one else says anything right away. 

Stu and Hannah have been listening, silently, wide-eyed.

Louis: “Jack – you should really talk about that first of all with Vivienne. You know that our labs have somewhat different approaches. And transfers get into all those questions of desk space and budgets. Anything’s possible, but Vivienne’s your go-to about your career path, OK?” 

Louis then claps his hands together, once.

That handclap is so interesting. What does clapping your hands together at that point signify in terms of the structure of the conversation? If it stands in for a spoken word or a sentence, what would that be? 

I have a guess, or a proposal. I think that if a single hand-clap plays the role of any word or utterance, it would be the word “Anyway”. It signals the end of a current topic, and either that a new topic is about to follow or, more likely, that we’re returning to the top-level topic that we were in before we digressed into the topic that we’re just now closing off. It pops the topic-stack, as it were. If I’m right then starting now we’re not going to be talking about the lab-transfer topic anymore tonight.

Louis: [Raising his voice.] “OK, everyone! Looks like the room is ready for us. Let’s head on over. They tell me that you don’t need to settle up at the bar, they’ll transfer everything over.”

Louis: [Not as loud, just to me] “Sorry we can’t invite you to the dinner itself, Jack. Each lab has its own morale budget. It’s great to see you socializing with my people though – we need to do more of that. We should get the teams together for another cross-lab thing, like we did – when was it? – in September? I’ll talk to Vivienne about it.”

People start grabbing coats from the back of their stools and picking up their drinks. Hannah gets off of seat #5 and brushes wordlessly by me toward the entrance. Stu follows her from seat #6 and chucks me on the shoulder as he passes – awkwardly, with his left hand, because he has a 3/4-full cocktail in his right hand. Soon it’s just me with my as-yet-untouched cocktail in front of me. Now what? Sounds like the bar bill is going to be transferred over to the private room for Louis’s people, so do I pay for mine because I’m not one of Louis’s people? No, I think – nobody knows, nobody cares, the bartender doesn’t know who is in what lab, and I’m sure that Louis’s so-called morale budget has enough slack in it to pay for one single Negroni or whatever this drink is. 

I realize that I’m really hungry. I check my phone – it’s 5:41pm. I’m half an hour later than my usual schedule, so no wonder. But there’s still plenty of time – Paesano’s serves the Chicken Marsala until 9pm.

——
Also republished on Medium