Hate Life, Will Travel

Derek Zumsteg, 10/2006
This story is licensed under a Creative Commons license, available here. You are, to sum up, free to copy, distribute, perform, fold, spindle, mutilate, or otherwise make any derivative work you wish as long as you attribute the work to me, the work is noncommerical, and derivative works using this license must be similarly licensed.

First, I'd like to thank the Yale SOM Behavioral Science Conference for inviting me to present. When we began this project years ago, we had no idea that we would attract so much attention, and I'm thankful for the opportunity to contribute to the field that has been my home ever since I realized I couldn't cut it in electrical engineering.

I'm kidding of course, I love the behavioral sciences. I'd like to go through what led me to head up this latest project, and discuss the initial field experiment we've carried out with the help of our sponsors. In case you forget who the project's sponsors are, you can look at the second projection screen here, which will rotate through them, or you can keep your eyes on me, as you'll note I'm wearing a T-shirt with similarly rotating advertisements. Also, instead of water I'll be drinking NOS energy drink throughout this presentation, even though it tastes like ass cut with battery acid. It's not that I'm without pride, it's that this project has proven ridiculously expensive and, unfortunately, teaching introductory sociology barely pays my own bills.

If you'll excuse me for just a second, I'll get started. Ahhh, that is delicious. And with an aftertaste so vile it makes you want to take another drink to wash it away. Anyway - building on the research by Dr. Torricelli of the University of Iowa - hey, you're here! I'm a huge fan -- can I buy you a drink later, sir? No, not one of these. Dear God, I don't teach at Pullman. Excellent, I'm so glad, we'll talk afterwards. Uh, sorry about that. Building on Dr. Torricelli's research, we identified about twenty five distinct gestures used by both men and women in flirtation rituals. We then catalogued variations, when they most frequently appear, and were able to discover some interesting correlations between the direction of gestures and cultural background, which we presented in the Journal of Human Interaction last year, but only because I didn't know who to blow at the International Behavior Sciences Journal. I kid, it's Dr. Foster.

The cataloging, though it revealed some interesting research question we divided among ourselves to pump up our publication count, didn't satisfy us. A list of words is only a dictionary. How do you put them together? What are the semantics, the linguistics of flirting? Both sides are attempting to attract a suitable mate, yeah yeah yeah, we've all taken psych 101 and then been shot down at the next party. How do words become speech? What's the difference between “how now brown cow” and “Now is the winter of our discontent, made something something…” I'm glad I wasn't an English major.

To this end, we attempted experimentation. Undergraduates were outfitted with a full-sensory recording rig and then instructed to turn them on when they began to flirt. This experiment failed to meet our expectations, for several reasons

1. Because the human interface labs get all the new gear, we had to make do with full sensory recording equipment several generations old. Students were forced to wear small over-ear box cameras with built-in microphones, and a backpack weighing about twenty pounds containing storage media and extended-use batteries
2. Because of the weight and unattractiveness of the equipment, we suffered a tremendous drop-out rate in our volunteer group
3. Undergraduates who did not drop out turned out to be disproportionately socially inept. We theorize that students more successful at flirtation saw the potential decrease in the effectiveness of their naturally-learned methods as not worth the extra credit we were able to offer
4. Students often did not recognize when someone was flirting with them
5. Because of the recognition problem, we required students to keep the equipment on at all time and self-edit flirtation, but found that they were extremely poor judges both of when and when they were and were not being flirted with
6. Also, it we quickly learned that the batteries overheated under constant use, subjecting students to discomfort and minor burns, and wearing the rigs resulted in injuries to the neck and lower back
7. People who attempted to flirt with socially inept undergraduates wearing bulky surveillance gear tended to themselves be poor flirters or people mocking the volunteer, thus rendering any results worthless for generalized application
8. In rare cases where the students wearing the equipment reached a successful conclusion, graduate students tended to horde the recordings, rather than check them into the archive

Still, we settled those lawsuits and somehow squeezed a paper out of those poor losers, and it us got into IBSJ - issue 1,100, if you haven't read it - and that was nice. We did not, however, get to go on the television and lecture circuits, or water down our research and write a best-selling pop science book, so we were forced to continue.

Sir, sir, please, sir, I don't mean you specifically, and I'm pretty sure that I do know who my father was. “The Life Beyond,” right? Great read, great read, it's always interesting to see how creative people can be when they're freed from the shackles of peer review. Can I ask you a quick question, though, Doctor Fisher? Do you have to buy carbon credits to offset all the hot air that comes out of your mouth? Thanks, I'll continue if that's fine with you.

Where was I - so we decided to try a different approach. We installed cameras and other surveillance equipment in the Harris undergraduate dorm common areas, and had graduate students monitor gatherings for flirtations, particularly those that were successful in producing a couple that left together, and map the flirtation gestures in order.

We encountered several problems with this experiment as well:

1. The sample size was small, consisting of one dormitory that houses approximately 500 students.
2. Graduate students tended to leave the office and join parties they deemed sufficiently interesting. Observer interference is a long-established side effect of scientific inquiry, and this was no exception.
3. When there were successful outcomes brought to fruition there in the common areas, the graduate students tended to keep the footage, rather than check it into the archive.

Despite those problems we found that flirtations follow a complicated but essentially finite decision tree, as you can see here. It's a very dense slide, I know. As you make a gesture, you move towards either a successful outcome, where you're mutually attracted and you go off and swap DNA, or an unsuccessful one, where either one or both of you decide that it's not a suitable match and break off communication. There are many variables involved, such as mood of both subject and flirtation target, setting, and each variable can change pathing, possible responses, how frequently each response is received, and so on.

For a variety of reasons, we made the difficult decision not to publish. However, we were able to take our unpublished results and use them to find additional grant money and begin the study which has brought you all here, unless you were told there would be refreshments and light snacks, in which case I should now mention that this is one of a series of experiments by doctors Frazier, Mogel, and Dawkins, on the effects of incentives on symposium attendance, and you can take up your entirely justified complaints with them. Sorry. I do have some delicious NOS energy drink samples up here, but they're not cold.

That's an interesting question. It seems like they'd be worse warm, but I don't see how that could be possible. Let me try one. Oh Dear God. I'm sorry. Let me just - does anyone want to go heat one of these up, we'll try for the cycle? Good.

In any event, the decision tree itself is deceptive. People don't loop through displaying their palm to the other person to demonstrate they're harmless over and over. It's a progression, and repetition is a failure to continue the ritual, so you're kicked out. In addition, even the best choice is not a guaranteed success - we found nearly identical paths, met with nearly identical responses ending in both failure and success, and there were no good predictors.

Further experimentation was needed. Our team was able to purchase modern full-sensory gear and wearable computers of sufficient power to support our next project. Our proposals also began to attract the attention of corporate sponsors interested in the nature of ritual of attraction, and with the project threatened by unexpected legal costs, we were happy to accept their help.

Which reminds me, I received word this morning that the Circuit Court of Appeals has rejected the class action lawsuit by Harris dormitory residents against us, affirming the state court's decision that our project was covered by both freedom of research and that the school's required consent to surveillance waiver was both sufficient in scope and legally binding. Our long legal nightmare is over, and I'm happy to announce that with the injunction lifted, we may now be able to publish our findings from that study.

Thank you. I'm excited, too. With footage? Sure, why not. $9.95 a month for access. No, no, of course we wouldn't do that. Well, we might seek additional funding through commercial licenses, but for scientific purposes, no, of course not.

Back to the program, though. In association with the Computer Sciences faculty, a combined team programmed a path-finding agent based on the known mappings from the previous experiment that could run on the new rigs. Now, a student seeking extra credit could plug in using a standard socket, easily concealed, and as an overlay see information about the target's gesture, and - here's the kicker - suggested responses, which would be based on the matrix we'd assembled except - and here's the other kicker - it would sometimes randomly pick new responses no one had chosen on the matrix. We'd get to see what they wanted to pick, what the influence of the system was, and as a kind of tax on them for giving them the advice, we'd do a little research at the same time.

The prompting system sucked at first, as you can see. I know, it's horrible, we didn't have any interface designers on the team. However, something unexpected happened. Beyond people continuing to show up to volunteer at the behavioral sciences lab, yes. It's the advantage of a research university, for those of you who aren't lucky enough: the laboratory mouse supply is entirely refreshed every four years.

Those who screened out as socially adept - the least likely to sign up for the first version of this experiment - were able to quickly reject bad prompts that would result inf failure, so they contributed faster pathing in generally productive directions. This allowed us to do some weighting of positions - and is where we began to map the transition boundaries.

At the other end, though, the undergrads who were most willing to take prompts were also the most forgiving of failure or strange prompts. In interviews they revealed that that if the flirtation failed, even if they had followed the most bizarre prompt possible, their belief was that they had screwed up. We got amazing data out of these kids.

So here's a sample progression, by a guy who rated himself as a 0 on a 0-10 scale in social aptitude on the volunteer survey.

Glance, glance
Smile, look away - okay, so she's coy, doesn't go out with everyone
Stretch, arch of brow - nice submissive gesture
Laughs too loudly, gets a lids down back
Swagger, hair flip - exposed neck, you see how this is going, so I'll skip ahead just a bit… fine, I won't skip ahead, it's grand gesture, lip lick, approach, talk, talk, smile, talk, talk, smile, more of the hip-swaying by both of them, he stands up real straight and then --
Accidental touch, return touch
Smile, smile
Beckon with finger, cocked head - we see curiosity
Lean in, lean in
Cheek-to-cheek talking annnnd they leave.

That's an almost optimal solution. We can attribute some of that to his confidence in the prompts, but still - I know, I know. Thank you. You're at a conference, feel free to use that yourself later, what happens in Montreal you know the rest.

What did we have? At every stage of the flirtation, every subject class gave us useful data. As they got more experienced, though, our map of known successful paths became pretty complete, and once subjects enjoyed success, they stuck to patterns they believed worked, so they were going through the same sequence more or less over and over. That wasn't real helpful for us, and it was okay that the original expert flirters dropped out over time.

When we went to recruit new subjects for a second round, we had more volunteers than we could possibly handle. We couldn't rotate the gear, either: even though they might ignore its advice, no existing subject wanted to give us back their devices. With massive demand and no supply, we had a problem in need of exploitation.

The team immediately formed a limited partnership, buying an exclusive development license to patents we'd filed that were owned by the university. We tried some licensing schemes for the program with map, spent a lot of money suing several former subjects into the ground for developing an open source version, but our top priority was getting as many of these into the hands of novices as possible, in order to map novice and intermediate paths as well as random ones, and also to find the best training methods and ways to successfully integrate our sponsors into common flirtation rituals.

Which reminds me. Ahhhh, that is some good energy drink. I'm wincing out of happiness. Enjoy NOS energy drink, folks, it's the best engine degreaser I've ever tasted.

We destroyed flirting at the University within a semester. Everyone wanted the answers to the test, but no one wanted the other side to be prompted, because giving false responses undermines the whole reason for flirting, this supposed exchange of secret, critical information about how healthy we are, how good a parent we'd be.

It also tainted our data, because once a prompter goes up against a prompter, all kinds of extremely strange things start to happen.

Yeah, that's a good question, and I don't want to get too specific, for reasons I'll get to, but I'll give you one. Prompters would unintentionally conspire. Novices would be asked to try something totally bizarre to attract initial attention, and other prompters, not having seen that behavior, would classify it as possible interest in their subject, and the person who did the weirdest thing at a party would get mobbed. But their prompter, having done something weird, would want to pick out the strangest return response as the most interesting, and direct their subject to respond to that… you see where I'm going.

No, I guess I can't guarantee that it was unintentional conspiracy, but they only dumped their data back to our archive every 72 hours, so real-time communication was impossible, and the data sets consisted only of system logs.

Okay, that might work, and that's a disturbing thought. Can we talk after? Thanks.

Two undergrads, both with prompters, went from smile exchange to leaving together in three gestures: head stretch from shoulder to shoulder, response of moving the head forward and backwards, stretching the neck, followed by a originally uncatalogued gesture made up of several trivial forms: interlaced fingers, turning palms up, thumbs towards the other, like so - while unlikely, it then became the optimal path for smile-to-sex. Reinforced by other prompters, pretty soon people not in the program knew what it meant, the sequence got a name - it had a lot of names, actually: Chicken Dance, the Double Fonzie, the Two-fer, the Cradle, I could go on for a while. Things went downhill from there. People discovered how to determine if the other person was being prompted and game them, there were parties designated as prompter-free, but then the reward for sneaking yours in was so high people did it anyway - chaos.

Four months after the invention of the three-gesture Chicken Dance, the primary means of flirting had turned to written communication and, surprisingly, American Sign Language - which our linguistics department wrote an excellent paper on, if you're interested. Everyone who had a prompter, had ever used a prompter, or acted like they had been prompter trained, all of them were toast.

And the social settings the prompters had learned dried up too - the clubs took a hit because we had dancing down, and even as they have recovered, it's not at all the same. Everyone who goes stands around and tries not to make eye contact. I understand it's like taking the subway.

Naturally, they all sued us: the volunteers, the people who bought or rented the device as a social aid, the club owners, people who were picked up by prompted people, people who weren't but had their social life disrupted, everyone. I can't say anything else about that on advice of my lawyers, and unfortunately I'm under court order to not discuss any version of the map - like this slide right here showing the last good map we generated before we broadened the distribution. I'm entirely forbidden to discuss this slide, but you're welcome to look at, cough cough, nudge nudge, my fellows, as I leave it up while I revive myself with a quick drink before I finish up.

Look at that, I accidentally displayed a slide I wasn't supposed to. I hope everyone managed to get some notes while I was quenching my will to live.

Yeah, so, the aftermath, sure. Everyone who started out being an expert flirter found themselves ostracized, along with everyone who was trained to become an expert flirter.

Ineptitude at flirting became expertise. If you spilled a drink on yourself, or you talked too fast, or any other of the hundred gaffes people make, it meant you were genuine, awkward, incompetent, nervous 100% human, and that meant you were extremely desireable.

Sense of humor, because we couldn't prompt jokes, and certainly not fast enough, and it's extremely hard to coach someone in comic timing, became as important as facial attractiveness, using the Ward and Harris evaluation.

The slick guys, if they were truly uniquely slick, continued to do well. But if you were generic, too similar to others, and the prompters learned your tricks, that was the end of you. Entire fraternities disbanded. No one wanted to go to their parties, no one wanted to date their members, which meant no member wanted to wear the letters… yeah, tragic loss, I know. There's a tear in my eye. I might be crying from the intense stomach cramps, though.

The nerds who didn't enlist or buy or rent a prompter from us… they're still making out like bandits. And not to give anything away, but I think we can simulate this for future versions if it doesn't wash out of the system in the next couple years, when hopefully we'll have settled this latest lawsuit and be able to market the devices again.

But we learned things - how swagger and self-confidence means that you're strong and can protect offspring, how grand gestures and playing around mean you demonstrate you're powerful, but won't hurt them, how showing off the best laptop demonstrates you have resources to provide for a family, and showing interest in others and having a sense of humor means that you'll be around tomorrow.

So what do you all think? Are there any questions? Does anyone want to finish my NOS?

-- DMZ, 2006