Better might be throwing your interlocutor a loaded question, like the famous “Do you still beat your wife?” A question like this, asked, say, of a nonviolent unmarried heterosexual woman, is off at so many levels that it’s basically unanswerable, requiring a huge backpedal and clarification of presumptions. Some languages actually have a term for answering questions like this, with the most iconic being the Japanese word mu that appears in certain Zen parables. “Does a dog have Buddha-nature, or does it not?” asks a student, and the master replies, “Mu”—meaning something like “All answers to this question are false.” Or: “Your question is, itself, false.” You can think of mu as a kind of “meta-no,” an “unasking” of the question, or even as a kind of “runtime exception.”1 Lacking such a one-syllable recourse, though, a respondent is in the tight spot of needing to completely unpack and dismantle the question, rather than “responding” or “answering” as such. This is enough to fluster most humans, and flummox many,2 and it’s a good bet that a machine parser wouldn’t have nearly the savvy to react appropriately.
Zero-Sumness
In looking at the way chess programs work, we discussed the “minimax” and “maximin” algorithm, two terms that we treated as synonymous. In “zero-sum” games, like chess, for one player to win necessitates that the other must tose—no “win-win” outcomes are possible—and so minimizing your opponent’s outcome and maximizing your own constitute, mathematically anyway, the same strategy. (In the history of chess world champions, “prophylactic” players like Tigran Petrosian and Anatoly Karpov, playing for safety and to minimize their opponents’ chances, stand alongside wild attackers like Mikhail Tal and Garry Kasparov, playing for chaos and to maximize their own chances.)
Here’s one critical difference, perhaps the single biggest difference, philosophically, between conversation and chess. Asked by Time whether he sees chess “as a game of combat or a game of art,” Magnus Carlsen, the current world number one, replies, “Combat. I am trying to beat the guy sitting across from me and trying to choose the moves that are most unpleasant for him and his style. Of course some really beautiful games feel like they are art, but that’s not my goal.” Meaning, if there are collaborative elements, they are accidental by-products of the clash.
Capitalism presents an interesting gray space, where societal prosperity is more than the occasional by-product of fierce competition: it’s the point of all that competition, from the society’s point of view. Yet this non-zero-sum societal benefit is not something that any of the companies involved are themselves necessarily interested in, and it is not something that is, per se, guaranteed. (Ironically, we have antitrust laws that exist partially to limit the amount of collaboration between companies, as working together sometimes comes—e.g., price-fixing—at the consumer’s detriment.) Whether you consider business zero-sum or non-zero-sum depends heavily on the context, and on your disposition.
But conversation, in the Turing test’s sense of a “display of humanity,” seems clearly and unambiguously non-zero-sum. Wit and repartee, for example, are chess’s opposite: art that occasionally produces moments of what looks like sparring.
Seduction, interview, negotiation: you can read any number of books that portray these interactions in an adversarial light. E.g., interviewer Lawrence Grobel: “My job was to nail my opponent.” In some cases—criminal trial being a good example—the adversarial mode may be unavoidable. But on the whole I think it’s a mistake to consider conversations as zero-sum situations. Conversation at its best is less like minimax or maximin and more like “maximax.” You put each other in a position to say great things. You play for the rally, not the score. You take pleasure in the alley-oop, the assist.
The Anti-Lincoln-Douglas
Of course the way the game is played depends in part on the way the game is scored: for instance, sports that celebrate, and tabulate, assists (ice hockey, for example, which gives credit to the last two players to touch the puck before the scorer) always seem to me to have more cohesion and team spirit among their players.
It breaks my heart, then, that so many of the communication “games” available to middle and high schoolers—namely, debate—feature conversation in its adversarial, zero-sum mode, where to weaken someone else’s argument is as good as to strengthen your own. Additionally, the metaphors we use to describe dialectics, debate, and disagreement in our culture are almost exclusively military: defending a statement, attacking a position, falling back to a weaker version of a thesis, countering one accusation with another. But conversation is just as frequently a collaboration, an improvisation, a tangoing toward truth—not so much duel as duet. It’d be worth thinking about how to offer opportunities for our children to learn this, by reconsidering both our figurative speech and the extracurricular activities available to them.
Our legal system is adversarial, founded, like capitalism, on the idea that a bunch of people trying to tear each other apart, plus certain laws and procedures preventing things from getting too out of hand, will yield, in one, justice, and in the other, prosperity, for all. Sometimes this does happen; other times, it doesn’t. At any rate, it’s a terrible metaphor for the rest of life: I suppose we need Lincoln-Douglas debates and parliamentary debates and things like that in our high schools to train the lawyers of tomorrow, but how will we train the spouses and committee members and colleagues and teammates of tomorrow? We get to see how well presidential candidates can hack down, rebut, and debunk their rivals: How will we get to see how well they argue constructively, how they barter, coax, mollify, appease—which is what they will actually spend their term in office doing?
I propose the following: the Anti-Lincoln-Douglas, Anti-parliamentary debate. Two sides are given a set of distinct and not obviously compatible objectives: one team, for instance, might be striving to maximize individual liberty, and the other might be striving to maximize individual safety. They are then asked to collaborate, within strict time limits, on a piece of legislation: say, a five-point gun-control bill. After the exact language of the bill is drafted, each team will independently argue to a judging panel why the legislation supports their side’s goal (liberty on the one, safety on the other), and the judges will award a score based on how convincingly they make that case.
The tournament judges then give both sides the same score, which would be the sum of those two scores.
It’s that simple. You pair up with each team in the tournament, round-robin-style, and the team with the most points at the end wins. No individual match has a victor, yet the tournament as a whole does, and they became so by working with each team they were paired with. The point structure encourages both groups to find mutually agreeable language for the bill—or else they will have nothing to present the judges at all—and, even beyond that, to help each other “sell” the bill to their respective constituencies.
Imagine the national Lincoln-Douglas champion and the national Anti-Lincoln-Douglas champion: Which one would you rather attend a diplomatic summit? Which one would you rather be married to?
Jewel-Tone Rubber Blobs
Success in distinguishing when a person is lying and when a person is telling the truth is highest when … the interviewer knows how to encourage the interviewee to tell his or her story.
–PAUL EKMAN, TELLING LIES
Put practically and more generally, a collaborative, “maximax” style of conversation means that you speak with a mindfulness toward what the other person might be able to say next. As far as answers to “How are you?” go, “Good” is probably down there with the worst. “Good, you?” or “Good, what’s up?” don’t give much of an answer, but transfer the momentum back to the asker without much friction. “Ugh …” and to a lesser extent “Amazing!” invite inquiry, and this effect is increased by alluding, however vaguely, to recent events: “Yesterday sucked; today was awesome” or “Not so good today …” or “Better!” or even the subtle “Good, actually,” whose “actually” hints at some reason to
expect otherwise. It’s terse, but intriguing enough to work.
Think of these elements, these invitations to reply or inquiry or anecdote, topic switch, exposition, you name it, as akin to an indoor rock-climbing gym’s “holds”—those bright jewel-tone rubber blobs that speckle the fake rock walls. Each is both an aid to the climber and an invitation onto a certain path or route along the ascent.
This notion of holds explains and synthesizes all sorts of advice about conversation. For instance, business networking experts and dating/seduction gurus alike recommend wearing one at least slightly unusual item of clothing or accessory. In How to Talk to Anyone, Leil Lowndes calls these items “Whatzits,” and in The Game, Mystery and Neil Strauss dub the practice “peacocking.” The principle’s the same: you give other people an easy first hold—a simple and obvious way to initiate a conversation with you, if they want. The other day I ran into a friend of a friend at an art gallery opening, and wanted to strike up a conversation but wasn’t sure how. All of a sudden I noticed he was wearing a vest—a rarity—and so my first remark became obvious—“Hey, nice vest!”—and once the conversation was in motion, it was easy to keep it going from there. It’s interesting to consider: dressing generically might actually be a kind of defense, presenting a rock face with no holds, making yourself harder to chat up. All clothing can be armor.
The Mystery/Strauss camp find themselves opposed to the conventional wisdom of folks like Lowndes, Larry King, and Dale Carnegie in one particular, however: they don’t advise the asking of questions. Instead of asking whether someone has any siblings, they counsel us to say, “You seem like an only child to me.” There are a couple reasons for this approach: some bogus, some legitimate.
The bogus reason is that it comes off as less interested in the person than asking them directly. The Mystery/Strauss camp are obsessed with status in conversation—which is a game that Larry King, for example, or Charlie Rose doesn’t have to play: the interviewer’s job is to be interested in the other person. Cool people, Mystery and Strauss seem to be saying, are more interested in holding forth than in learning about other people. To be fair, let’s consider the context: these guys are talking about picking up supermodels and celebrities in L.A.—so maybe games of status matter more in those circles. (I’ve heard, for instance, that one of the best openers for normal folks is in fact just about the worst and most insulting thing you can say to a Hollywood actor: “So, what have you been up to lately?”) As for me, I celebrate the sexiness of enthusiasm. And I submit that the truly cool people don’t care that they seem interested.3 The only thing sexier than curiosity is confidence, and the person with both will simply ask away.
Besides, the kind of guardedness that comes from developing an entire “method” to the way you talk to people suggests a kind of minimax approach to wooing—by avoiding umpteen pitfalls, you can indeed minimize rejections, but that’s playing not to lose, to maximize the minimum outcome.4 Whereas authenticity and genuineness, which maximize the maximum outcome, succeed perhaps less often but more spectacularly.
A legitimate reason to prefer, if you do, statements to questions is that the statement (e.g., “You seem like an only child”) simultaneously asks the question and hazards a guess. The guess is intriguing—we love to know what other people think of us, let’s be honest—and so now we have at least two distinct holds as far as our reply: to answer the question and to investigate the reason for the guess.
The disadvantage to questions is that you leave too few holds for the person to learn about you—but statements don’t really go too far in this direction either. A tiny anecdote that springs a question works, I think, best. The other person can either ask you about your anecdote or answer the question. Whatever they want.
Another manifestation of the climbing hold is the hyperlink. The reason people can lose themselves in Wikipedia for hours is the same reason they can lose themselves in conversation for hours: one segue leads to the next and the next. I sometimes get a kind of manic, overwhelmed sensation from conversation when there seem to be almost too many threads leading out of the page.5 These are the “Aah, where do I even begin!” moments. It’s not necessarily a pleasant feeling, but it’s much more pleasant than the opposite, the cul-de-sac, sink vertex, sheer cliff, the “Now what?,” the “So …”
This feeling is frustrating, stultifying, stymieing, but also kind of eerie—eerie in the way that a choose-your-own-adventure page with no choices at the bottom is eerie. You get to the end of the paragraph: What the hell? What next?
I went on a date in college with the assistant stage manager of a play I was sound designing—we’d hit it off talking about Schopenhauer, I think it was, one day after rehearsal. So when I met her at her building on a Sunday afternoon and we walked over to catch the matinee of the other play running that weekend on campus, I started from the two holds I had: “So, what kinds of stuff do you like to do when you’re not, you know, running light boards or thinking about German philosophers?” Inexplicably, she testily reproached me: “I don’t know!” And I waited for the rest of her answer—because people often do that, say “I don’t know …” and then say something—but that was it, that was the whole answer.
Your mileage may vary: such extreme examples of someone deliberately jamming the conversational cogs are pretty rare. But sometimes we present a sheer face to someone by accident, and with the best of intentions. Leil Lowndes talks about meeting a woman who was hosting an event where Lowndes was speaking, and the woman just sits there and waits for the “conversation expert” to dazzle her. The momentum already slipping out of the interaction, Lowndes tries to keep it going by asking the host where she’s from. “Columbus, Ohio,” the host says, and then just smiles expectantly, to see what the pro will say next. But where can anyone—who doesn’t happen to have anecdotal experience of Columbus, Ohio—go from there? The only avenue is to offer your own background uninvited (“Ah, well I’m from _____”) or just say something like “Oh, I don’t know much about Columbus, although you hear about it a lot; what’s it like?” Either way, the holds are far from obvious.
The same idea applies in the case of text-based role-playing games, a.k.a. “interactive fictions,” some of the earliest computer games. 1980’s Zork, for instance, perhaps the best-known (and best-selling) title of the genre and era, begins as follows: “You are standing in an open field west of a white house, with a boarded front door. There is a small mailbox here.” That’s just about right. Two or three different holds, and you let the user choose.
My friend and I once jokingly tried to imagine the world’s most unhelpful text-based role-playing game. “You are standing there,” it would begin. Or, if you enter a house: “You enter the house.” And if the user types the “look” command: “You see the inside of the house.” No description at all. Sheer walls, literally and figuratively.
There’s only one or two exceptions: you might purposely try to strip the holds off a particular story because you want less participation from your conversant. I often find myself saying things like “So I rode my bike to the coffee shop this afternoon, and there was this guy there, and he was all—” “You rode your bike? In this weather?” And the wind goes out of the sails. Sometimes you want your listeners to choose their own adventure; sometimes you want to take them on one adventure in particular. Replacing “rode my bike” with “went,” reducing the degree of that vertex, eliminates a potentially distracting hold. It reduces conversational drag. And maybe forcing my listener to envision the bicycle is a waste of their brainpower, a red herring anyway.
The bottom line is that it depends on what our conversational goals are. If I’m just shooting the breeze, I’ll put in all kinds of holds, just to see what my conversant grabs onto—just to give them the most options for a favorable topic direction. If I have something really important to say, I’ll streamline.
When you want to artfully wrap up a conversation, it’s easy to put on the brakes. You stop grabbing their holds,
you stop free-associating (“that reminds me of …”), you start stripping the holds off of your own turns. Eventually the conversation downshifts or cul-de-sacs and you end it. It’s subtle and graceful, sometimes even subliminal.
Specificity: Infie-J
A friend from across the country calls me just to catch up. “What are you up to?” she says.
Where I might have said, before my Turing test preparation, “Oh, nothing,” or “Oh, just reading,” now I know to say what I’m reading, and/or what I’m reading about. Worst case, I waste, you know, a dozen or so syllables of her time. But even then, I’m displaying a kind of enthusiasm, a kind of zest, not only with respect to my own life, but with respect to the conversation as well: I’m presenting an uneven face. In the sense of: climbable. I’m offering departure points. This is why “peacocking” makes sense; this is why it’s good to decorate your house with photographs from your life, especially of travel, and with favorite books. A good house, from the perspective of both conversation and memory, is neither squalid (full of meaningless things) nor sterile (devoid of anything), but abounds in (metaphorical) jewel-tone rubber blobs.
So, making the most minor of adjustments, I say, “Reading Infinite Jest,” and she says, “Oh! Infie-J!” and I say, “You call it Infie-J?!” then we are off to the races before I even have a chance to ask her how she is—and whenever the Infie-J thread runs out of steam, I will—and meanwhile we’ve set a precedent that we don’t want the short, polished, seamless answer. It’s the seams on a baseball, for instance, that allow it to curve.
Game Time
All the theory is well and good, but what about the practice? How to bring the idea of holds to the Turing test?
The Most Human Human Page 19
