 I hate playing escape rooms with that one guy who just absolutely refuses to ask for a clue, even though everybody else in the group is clearly frustrated and rolling their eyes. It's like, come on man, take a hint. In 2007, a meta-filter user posted a rant about how a dissent acquaintance had asked if she could stay in his family's small apartment while she was visiting his city for work. He expressed frustration at the presumption of making such a request without any hint that it might be okay, and petitioned other forum members for help in crafting a polite excuse to get out of hosting her. Some responded with commiseration for his predicament, but others expressed confusion. I mean, she was asking, wasn't she? Surely if his answer was no, he could just say no. One user, Tangerine, observed that these responses seemed to stem from two distinct understandings of how favors ought to work, terming them guest culture and ask culture. In some situations, we only ask someone for a favor once it's been fairly well-signaled that they're probably amenable to it, once we can guess that they'll say yes. In other situations, we feel comfortable asking for something bluntly, expecting that if the person we're asking doesn't feel up to it, they'll just say no. If you're having dinner with a partner, it's usually totally fine to ask if they want to split entrees, and it's totally fine for them to turn you down. But if you get down on one knee in front of a group of friends and ask them to marry you with the same uncertainty as to their response, you're probably going to have a bad time. Different people have different standards of what sort of situations warrant guessing or asking, and when there's a mismatch, as in the Metafilter example, it can suck for everyone involved. Where those thresholds are set depends a lot on what sort of culture you live in. Low-context cultures, like the U.S. and Germany, tend to focus on direct and unqualified communication, where the words used in this statement are all you need to know to understand its meaning. In a low-context culture, you might ask to stay at someone's place and get a response like, absolutely not, you snore and it's annoying. Ouch. Not likely to make you popular with anyone listening, but literally true, and you did ask. High-context cultures, on the other hand, as seen in China or Africa, tend to pull in the shared history and social status of the people interacting while they're communicating, relying on subtle cues conveyed by norms and implication, often to save face. Depending on the elements surrounding what's being said, the explicit language used might not be important at all. In a high-context culture, you might never mention staying over at someone's place, but they recognize your predicament and invite you. You poetly decline. They invite you again, saying it would be rude of you to refuse, so you finally agree, and they're upset with you for accepting. You're probably from a lower-context culture, so that amount of social theater might sound agonizingly inefficient, but even in the lowest-context cultures, we follow similar norms of indirect communication, hinting to someone that their fly is down instead of announcing it to the room, offering a glass of water to someone who looks especially drunk. There are at least two ways of interpreting these phenomena. Maybe all these polite flourishes are just frilly nonsense, diluting the meaning of speech and wasting our precious time by failing to get to the point. Maybe we've all been languishing in an unnecessary mire of confusing inanity and face-saving indirectness when we could be getting things done with more accurate speech. On the other hand, maybe there's something useful happening outside of raw information transfer, something that justifies those seeming inefficiencies. In their paper, The Logic of Indirect Speech, Stephen Pinker and colleagues identify some notable strategic advantages to communicating in less explicit ways, which might help to explain why every culture seems to normalize at least some non-explicit speech. They build their model on a simple game-theoretic scenario, a coordination game, where the goal is to correctly predict and synchronize one's decisions with another player. Say you've been pulled over for speeding, and a police officer is in the process of writing you a ticket that you'd rather not have. It's possible that the officer would accept a bribe to let you off the hook, but it's also possible that he's a narc who'll penalize you even more for your breach of protocol. At first glance, your options aren't rosy. Either stay quiet and get stuck with the ticket, or take a huge gamble on the off chance that he'll take a bribe and let you go. But what if there was another way? What if you could communicate your desire to bribe the officer only if he might be interested? In this situation, indirect speech is a strategy that has better outcomes than either of the direct speech options. Instead of saying, will you let me off with a warning for 60 bucks? Maybe something like, is there anything we can do here so I don't get a point on my record? Well, even an honest driver uninterested in a bribe might make that sort of statement, and the threshold for acting on it, either by taking the bribe or calling you on it without hard evidence, is different for straight and crooked cops. In a nutshell, the right kind of indirect speech gives you a leg up in your coordination game by changing the game for only a certain kind of player. But that doesn't really seem like it characterizes all indirect speech. It's not like hinting that your roommate should clean up the kitchen is an encoded message intended only for sympathetic roommates. Actually, the paper's author suggests that it's the same basic coordination game, but referred to a model known as politeness theory to explain what you're playing for. In this model, speech has two distinct goals, conveying the literal informational content of the words being said, and signaling certain social relationships between the people talking, things like dominance, communality, or reciprocity, each of which has a unique set of associated values and linguistic norms. Use a speech pattern for one relationship type in the wrong context, and it causes awkwardness. If the new hire at work starts bossing coworkers around, there's a mismatch between the expected social role of coworkers, reciprocity, and the role they're exhibiting, dominance. If you ask a friend to help you move and they demand payment for services rendered, they're subverting their established communal role in favor of reciprocity, and it's weird. In situations where social relationships aren't crystal clear, or a statement of a particular type might be construed as awkwardly violating an established role, indirect language plays similarly as in the bribing situation. It allows the speaker a third option, with a higher payoff than staying quiet, and with less risk than just saying something directly, although this time the risk is of impropriety and awkwardness instead of jail time. If you hint to your roommate that the dishes might stand washing, you can communicate the desire for them to do it without potentially coming off as a bossy boots. If your friend agrees easily to your request for moving help, and obliquely mentions that it's going to be a fair amount of gas money to get it done, you can maintain your communal relationship and still fork over some cash in a reciprocal fashion. But what about those situations where there's no uncertainty whatsoever about what's being implied? Or someone's communicating a very clear message, just using different words? The game theory model might work when there's some amount of reasonable uncertainty as to your intent, but if everyone knows exactly what's being said anyways, what's the point? Well, the paper's authors note that we tend to think of language as an all-or-nothing phenomenon. Either we say something, or we don't. Either it's true, or it's not. That perception makes indirect speech a useful tool even when it's obvious what's being said, or, you know, not said. We talked a little about common knowledge back in episode 150, the idea that there's a huge difference between everyone knowing a thing and everyone knowing that everyone knows it. The slightest theoretical possibility that someone might not have put two and two together for some bit of indirect speech circumvents that common knowledge element. After all, nobody's said it out loud yet. Maybe it went over someone's head. Because of our perception of language as being such a binary thing, that allows us to create shared personal knowledge without making it common knowledge, which, in turn, allows certain exchanges without social consequence. If I bluntly ask, could you share this episode with your friends, and you say, no, sorry, it's not that good. There's a sting to that. Everyone who saw that interaction would be like, oh, snap, Josh just got shot down. On the other hand, if I say, you know, the share button's right down there, and you play dumb, like, oh, thanks for that. I know they moved it around a bit in the last redesign. Why do they keep doing that? It's perfectly clear to anyone with half a brain what just happened, but it doesn't have the same social fallout that it would in the first case. After all, nothing was explicitly said, and if someone wasn't paying attention, it might have gone unnoticed, so we can just play it off like it didn't happen. It's a way to preserve harmony in situations that would otherwise demand some sort of reshuffling of social roles and relationships. That tiny sliver of plausible deniability can be seen operating in high-context cultures and they're disdained for direct confrontation. It's very rare that anyone has to deal with the social consequences of being told off. With the observations raised in this paper, it becomes clear why that original metafilter poster might feel uncomfortable and frustrated, with an acquaintance boldly asking to stay at his place without any indirect hinting or implication that it might be okay. She's not offering any sort of payment or exchange for the favor, so she's appealing to a relationship of commonality, which he obviously doesn't feel as appropriate. Awkward. She's asking explicitly, which means that he has to work extra hard to come up with a suitably face-preserving way to say no. Sure, it's a more straightforward transaction, and if her only concern was getting a fast yes or no so she could get on with planning her trip, maybe that's all it needed to be, but she can probably bet on getting a lot more nos than if she tried asking more indirectly, especially with guest culture folks. So, indirect speech provides numerous strategic advantages over explicit, straightforward speech in some situations, even if it's less efficient. That's not to say that it's always warranted or that there's no benefit to directness. For many people, trying to puzzle out what's not laid bare is frustrating, maybe even impossible, but understanding the benefits of well-executed subtlety might make it a little less frustrating when someone won't come right out and say something. Do you find the logic of indirect speech a compelling enough justification for communication that might appear simply inefficient? Did you know that there's a Thunk Patreon and Discord server? Please leave a comment below and let me know what you think. Thank you very much for watching. Don't forget to… Okay, come on back. I've got one more thing. I've been trying for a week now to find some sort of literature that directly tests the idea, but decision theory research is over here, and linguistics research is over there, and sociology research is in the third place. It's just very hard to find what I'm looking for. I have another possible explanation for the ubiquitous nature of indirect speech that isn't about strategic advantage for the individual using it, but more about effective coordination of groups. In computer modeling, there are certain questions that are just too hard to calculate directly. Take the traveling salesman problem. Given a set of interconnected towns, find the shortest complete loop that goes through each town only once. The computational complexity of that problem scales insanely quickly. Once you have more than a handful of towns, the universe will burn itself out before any supercomputer could get close to finding an answer. Among these sorts of impossibly difficult computation problems are multi-dimensional optimization problems. Given a number of individuals, each with a handful of discrete preferences and priorities, possibly including the satisfaction of other individuals, finding the one decision that results in the greatest possible satisfaction for everyone is simply impossible. But researchers have found sneaky ways to get around these computationally intractable problems, using a technique called relaxation. Instead of trying to solve the problem exactly as stated, they calculate answers of simpler related problems. Relaxing the constraints slightly to make the result exponentially easier to compute. Rather than trying to brute force our way through the traveling salesman problem, what if we allow backtracking, visiting towns more than once with a small penalty for cheating? All of a sudden, the problem becomes fantastically easy. Computers can return results in milliseconds rather than eons. The results aren't precisely what we're after, of course, but they provide substantial insight into the original problem, defining upper and lower bounds on what the exact answer might be. The same principle could be applied to multi-dimensional optimization problems. By relaxing discrete preferences into fuzzy suggestions, rather than finding the single truly optimal value of the problem, sometime after the sun goes supernova, we can get an answer that's probably pretty close to our theoretical maximum almost instantly. That approach bears a remarkable resemblance to the indirect dancing people do in groups when they're trying to reach a consensus. They hint at their own preferences. They quickly back away from suggestions that don't seem to resonate with the group. They search for subtle indications that others might be thinking what they're thinking. Everyone's desires exist in a vague probability cloud rather than explicit statements. When someone comes right out and says, I want to do this, it's not uncommon that people will sigh in frustration rather than welcoming it as a helpful data point. Now, the fuzzy convergence they were working towards has to happen around this discrete constraint of one person's overt preference. And that process is hamstrung by absolutely needing to take that preference into account in their solution, which will now probably take longer. Anyways, I haven't been able to find either a confirmation or a refutation of this concept in the literature. I have a buttload of papers linked in the video description, if you want to try your hand at chasing down something more concrete. Or, if you're a grad student in one of those fields looking for a pretty cool cross-disciplinary thesis, I put together some citations for you. Either way, I'd welcome any commentary on the idea. Thank you very much for watching. Don't forget to blah, blah, subscribe, blah, share, and don't stop thunking.