The most commonly used introduction to signaling, promoted both by Robin Hanson and in The Art of Strategy, starts with college degrees. Suppose, there are two kinds of people, smart people and stupid people; and suppose, with wild starry-eyed optimism, that the populace is split 50-50 between them. Smart people would add enough value to a company to be worth a $100,000 salary each year, but stupid people would only be worth $40,000. And employers, no matter how hard they try to come up with silly lateral-thinking interview questions like “How many ping-pong balls could fit in the Sistine Chapel?”, can't tell the difference between them.
Now suppose a certain college course, which costs $50,000, passes all smart people but flunks half the stupid people. A strategic employer might declare a policy of hiring (for a one year job; let's keep this model simple) graduates at $100,000 and non-graduates at $40,000.
Why? Consider the thought process of a smart person when deciding whether or not to take the course. She thinks “I am smart, so if I take the course, I will certainly pass. Then I will make an extra $60,000 at this job. So my costs are $50,000, and my benefits are $60,000. Sounds like a good deal.”
The stupid person, on the other hand, thinks: “As a stupid person, if I take the course, I have a 50% chance of passing and making $60,000 extra, and a 50% chance of failing and making $0 extra. My expected benefit is $30,000, but my expected cost is $50,000. I'll stay out of school and take the $40,000 salary for non-graduates.”
...assuming that stupid people all know they're stupid, and that they're all perfectly rational experts at game theory, to name two of several dubious premises here. Yet despite its flaws, this model does give some interesting results. For example, it suggests that rational employers will base decisions upon - and rational employees enroll in - college courses, even if those courses teach nothing of any value. So an investment bank might reject someone who had no college education, even while hiring someone who studied Art History, not known for its relevance to derivative trading.
We'll return to the specific example of education later, but for now it is more important to focus on the general definition that X signals Y if X is more likely to be true when Y is true than when Y is false. Amoral self-interested agents after the $60,000 salary bonus for intelligence, whether they are smart or stupid, will always say “Yes, I'm smart” if you ask them. So saying “I am smart” is not a signal of intelligence. Having a college degree is a signal of intelligence, because a smart person is more likely to get one than a stupid person.
Life frequently throws us into situations where we want to convince other people of something. If we are employees, we want to convince bosses we are skillful, honest, and hard-working. If we run the company, we want to convince customers we have superior products. If we are on the dating scene, we want to show potential mates that we are charming, funny, wealthy, interesting, you name it.
In some of these cases, mere assertion goes a long way. If I tell my employer at a job interview that I speak fluent Spanish, I'll probably get asked to talk to a Spanish-speaker at my job, will either succeed or fail, and if I fail will have a lot of questions to answer and probably get fired - or at the very least be in more trouble than if I'd just admitted I didn't speak Spanish to begin with. Here society and its system of reputational penalties help turn mere assertion into a credible signal: asserting I speak Spanish is costlier if I don't speak Spanish than if I do, and so is believable.
In other cases, mere assertion doesn't work. If I'm at a seedy bar looking for a one-night stand, I can tell a girl I'm totally a multimillionaire and feel relatively sure I won't be found out until after that one night - and so in this she would be naive to believe me, unless I did something only a real multimillionaire could, like give her an expensive diamond necklace.
How expensive a diamond necklace, exactly? To absolutely prove I am a millionaire, only a million dollars worth of diamonds will do; $10,000 worth of diamonds could in theory come from anyone with at least $10,000. But in practice, people only care so much about impressing a girl at a seedy bar; if everyone cares about the same amount, the amount they'll spend on the signal depends mostly on their marginal utility of money, which in turn depends mostly on how much they have. Both a millionaire and a tenthousandaire can afford to buy $10,000 worth of diamonds, but only the millionaire can afford to buy $10,000 worth of diamonds on a whim. If in general people are only willing to spend 1% of their money on an impulse gift, then $10,000 is sufficient evidence that I am a millionaire.
But when the stakes are high, signals can get prohibitively costly. If a dozen millionaires are wooing Helen of Troy, the most beautiful woman in the world, and willing to spend arbitrarily much money on her - and if they all believe Helen will choose the richest among them - then if I only spend $10,000 on her I'll be outshone by a millionaire who spends the full million. Thus, if I want any chance with her at all, then even if I am genuinely the richest man around I might have to squander my entire fortune on diamonds.
This raises an important point: signaling can be really horrible. What if none of us are entirely sure how much Helen's other suitors have? It might be rational for all of us to spend everything we have on diamonds for her. Then twelve millionaires lose their fortunes, eleven of them for nothing. And this isn't some kind of wealth transfer - for all we know, Helen might not even like diamonds; maybe she locks them in her jewelry box after the wedding and never thinks about them again. It's about as economically productive as digging a big hole and throwing money into it.
If all twelve millionaires could get together beforehand and compare their wealth, and agree that only the wealthiest one would woo Helen, then they could all save their fortunes and the result would be exactly the same: Helen marries the wealthiest. If all twelve millionaires are remarkably trustworthy, maybe they can pull it off. But if any of them believe the others might lie about their wealth, or that one of the poorer men might covertly break their pact and woo Helen with gifts, then they've got to go through with the whole awful “everyone wastes everything they have on shiny rocks” ordeal.
Examples of destructive signaling are not limited to hypotheticals. Even if one does not believe Jared Diamond's hypothesis that Easter Island civilization collapsed after chieftains expended all of their resources trying to out-signal each other by building larger and larger stone heads, one can look at Nikolai Roussanov's study on how the dynamics of signaling games in US minority communities encourage conspicuous consumption and prevent members of those communities from investing in education and other important goods.
The Art of Strategy even advances the surprising hypothesis that corporate advertising can be a form of signaling. When a company advertises during the Super Bowl or some other high-visibility event, it costs a lot of money. To be able to afford the commercial, the company must be pretty wealthy; which in turn means it probably sells popular products and isn't going to collapse and leave its customers in the lurch. And to want to afford the commercial, the company must be pretty confident in its product: advertising that you should shop at Wal-Mart is more profitable if you shop at Wal-Mart, love it, and keep coming back than if you're likely to go to Wal-Mart, hate it, and leave without buying anything. This signaling, too, can become destructive: if every other company in your industry is buying Super Bowl commercials, then none of them have a comparative advantage and they're in exactly the same relative position as if none of them bought Super Bowl commercials - throwing money away just as in the diamond example.
Most of us cannot afford a Super Bowl commercial or a diamond necklace, and less people may build giant stone heads than during Easter Island's golden age, but a surprising amount of everyday life can be explained by signaling. For example, why did about 50% of readers get a mental flinch and an overpowering urge to correct me when I used “less” instead of “fewer” in the sentence above? According to Paul Fussell's “Guide Through The American Class System” (ht SIAI mailing list), nitpicky attention to good grammar, even when a sentence is perfectly clear without it, can be a way to signal education, and hence intelligence and probably social class. I would not dare to summarize Fussell's guide here, but it shattered my illusion that I mostly avoid thinking about class signals, and instead convinced me that pretty much everything I do from waking up in the morning to going to bed at night is a class signal. On flowers:
Anyone imagining that just any sort of flowers can be presented in the front of a house without status jeopardy would be wrong. Upper-middle-class flowers are rhododendrons, tiger lilies, amaryllis, columbine, clematis, and roses, except for bright-red ones. One way to learn which flowers are vulgar is to notice the varieties favored on Sunday-morning TV religious programs like Rex Humbard's or Robert Schuller's. There you will see primarily geraniums (red are lower than pink), poinsettias, and chrysanthemums, and you will know instantly, without even attending to the quality of the discourse, that you are looking at a high-prole setup. Other prole flowers include anything too vividly red, like red tulips. Declassed also are phlox, zinnias, salvia, gladioli, begonias, dahlias, fuchsias, and petunias. Members of the middle class will sometimes hope to mitigate the vulgarity of bright-red flowers by planting them in a rotting wheelbarrow or rowboat displayed on the front lawn, but seldom with success.
Seriously, read the essay.
In conclusion, a signal is a method of conveying information among not-necessarily-trustworthy parties by performing an action which is more likely or less costly if the information is true than if it is not true. Because signals are often costly, they can sometimes lead to a depressing waste of resources, but in other cases they may be the only way to believably convey important information.
Be warned! Signaling that you understand signaling is a terrible signal, because it throws all your other signals into doubt. Revealing that you are optimizing your signaling separately (for example, talking about "PUA") is among the worst signals of all.
This is the opposite of true. People want allies that are competent signalers. Explicitly talking about signalling is in most cases a bad idea and usually a signal that you don't understand signalling or respect it sufficiently.
To the extent that there are "worst signals of all" rather than signals being dependent on context and goals, talking about PUA wouldn't be near the top. There are more than enough cases where it is either neutral or positive---especially when that ridiculous acronym isn't used.
What you're saying rings true, and a lot of people agree with you, but is it actually right? Is it testable? I can think of plenty of counterexamples, by people who look like they know they're doing. But I can't think of anyone whom I just want to grab and yell at: "you'd be so effective if you'd just shut up about the signaling already!"
I enjoyed this post.
It also hints at the notion of signaling equilibria. Consider the Helen of Troy example - this is clearly not an equilibrium, because Helen ends up marrying a bankrupt. Soon "spends lots of money on diamonds" will no longer be a signal of wealth, but will instead be a signal of profligacy - as indeed it is where I live. A man walking around in flashy jewellery would be considered low-class, presumably because in the past there has been exactly this signaling reversal.
In a stable signaling equilibrium, the signal needs to be hard-to-fake. This is why easy-to-fake signals are unstable - in the flowers example, the proles can and will catch on, and switch to the upper-middle-class flowers, so the upper-middle-class have to keep moving to stay ahead of them. The same phenomenon is seen in baby names, where upper-middle-class names become prole after a generation.
One thing I would have preferred is a discussion of the positive externalities of signaling, not just the negative ones. For example, if Yvain and lukeprog are both trying to signal their superior intelligence by writing insightful posts, this may get into an "arms race" for them, losing utility. However, the Lesswrong community gains utility overall. I think the externalities of signaling are generally positive in the real world, they only tend to be negative in what are anyway zero-sum games (e.g. begging).
You have just summarized "civilisation" in a nutshell.
A simple, interesting, complementary fact is that the cigarette manufacturers all saw profits skyrocket when laws started banning cigarette ads on TV. All of their products are largely interchangeable, so advertising doesn't tell you anything new about the product, it just builds brand loyalty. So it saves everyone costly signalling.
It's also extremely difficult for new cigarette manufacturers to break into the market. It's very hard to use a really clever ad campaign to increase your market share when you're not allowed to advertise on TV. Curiously, this may actually harm consumers, in that it prevents competition from lowering cigarette prices. I suppose this analogizes to the idea that if everyone were suddenly banned from displaying their wealth, it would be very difficult to woo Helen of Troy unless you had clearly shown your wealth prior to the ban. Thus, banning signalling can lead to losses, as the wealthiest suitor may be unable to woo Helen if he came to the game too late.
That's interesting. Do you have a cite?
I had long ago (but after being heavily influenced by Overcoming Bias) thought that signaling could be seen simply as a corollary to Bayes' theorem. That is, when one says something, one knows that its effect on a listener will depend on the listener's rational updating on the fact that one said it. If one wants the listener to behave as if X is true, one should say something that the listener would only expect in case X is true.
Thinking in this way, one quickly arrives at conclusions like "oh, so hard-to-fake signals are stronger" and "if everyone starts sending the same signal in the same way, that makes it a lot weaker", which test quite well against observations of the real world.
Powerful corollary: we should expect signaling, along with these basic properties, to be prominent in any group of intelligent minds. For example, math departments and alien civilizations. (Non-example: solitary AI foom.)
Possibly a side issue, but one motivation for signalling occurs when measurement is difficult for some reason e.g. regulation.
Giving prospective employees an IQ test can be quite hazardous for the employer. http://en.wikipedia.org/wiki/Intelligence_and_public_policy
Spending $50,000 on college - and incurring toxic student debt in the process - to prove something that can be demonstrated by a $500 test seems strange, in the absence of other factors. Particularly when colleges use a near-IQ test (SAT) as one important criterion for admission!
College degrees are better signals for conscientiousness than intelligence, which is no coincidence, since employers in real life care more about conscientiousness.
I doubt this is true. I've seen research that claims that on average, college students spend less than an hour a day studying. I've attended 3 universities in my life (undergrad to grad school), and skipping classes frequently, dressing like a slob in class, and skipping the required reading seem typical. If I cared mostly about conscientiousness, I would be more impressed by someone holding down a job at McDonald's for 4 years than graduating college, because a McDonald's manager has no problem with firing someone who skips work frequently. Most college professors don't even take attendance.
Yet long-term McDonald's employees get very little career boost from this in applying for jobs at Goldman Sachs or whatever. A kid who manages an Art History degree at Harvard while mostly partying and doing the minimal work to pass has a vastly better chance than the a long-term McDonald's employee with a sterling letter of recommendation from his boss.
This is vastly over-simplified. I did an internship at a firm that designs ... (read more)
Using the term over-simplified was my attempt at generosity. As presently stated, your claim is entirely wrong. Intelligence is the single best predictor of job performance for all but the most narrowly-focused manual tasks, see for example Ree & Earles, Current Directions in Psychological Science vol. 1, No. 3 (Jun., 1992), pp. 86-89.
The strong claim you made in your original comment was entirely false, and I get the impression you were just speculating wildly about something you don't actually know much about.
After intelligence, Conscientiousness is probably the single best predictor of job success since it predicts even after controlling for IQ, education level, etc. (Cribbing from my usual footnote, the best starting point is the meta-analysis http://people.tamu.edu/~mbarrick/Pubs/1991_Barrick_Mount.pdf )
Conformity too. This is a factor often overlooked in discussions of this sort.
(There are in fact two ways in which education signals conformity. The first one is the fact that you have conformed to the social norm that you are supposed to signal your intelligence and conscientiousness with this particular costly and wasteful endeavor, not in some alternative way that would signal these traits just as well. The second one is that you have successfully functioned for several years in an institution that enforces an especially high level of conformity with certain norms of behavior that are especially important in a professional context.)
I read a George Will column where he said that aptitude tests for jobs used to be much more common. Then some employers got sued because racial minorities were doing too poorly on the tests. Requiring a college degree is legally safer than an aptitude test (even though there is at least as much racial disparity in college degrees as aptitude tests).
If you want to be a police officer in New London, Connecticut, you'd better not score too high on the IQ test.
A new perspective on rational charities and inefficient fuzzy-purchasing:
Donating to charity is an expensive signal of high moral character or status. Enough people have caught on to this signalling method that in order to signal with your donations now, you have to donate to some 'conspicuous consumption'-style charity (kittens with rare diseases, receiving visits from their rockstar idols).
Except in real life, the #1 signaling college has a graduation rate of 98% and an average Grade Point Average of A-.
As the saying goes, the "only way to flunk out of Harvard is to die of a heroin overdose."
I have often wondered if anyone has tried to save their acceptance letters from colleges they couldn't afford to go to and show them to employers. Why doesn't this work?
When I was studying under Amotz Zahavi (originator of the handicap principle theory, which is what you're actually discussing), he used to make the exact same points. In fact, he used to say that "no communication is reliable unless it has a cost".
Having this outlook on life in the past 5 years made a lot of things seem very different - small questions like why some people don't use seatbelts and brag about it, or why men on dates leave big tips; but also bigger questions like advertizing, how hierarchical relationships really work, etc.
Also explained a lot about possible origins of (some) altruistic behaviors; Zahavi's favorite examples were from the research he and his wife conducted, wherein they observed small groups of social birds (forgot the species, sorry) where altrusitic behavior is common. And it turns out, it's the dominant birds who behave altrusitically, rather than exploit their weaker brethren - but doing so as a show of strength. My own favorite example is when a lower-status male caught a worm and tried feeding it to the alpha male. The latter proceeded to beat him up, take the morsel, and force-feed it back to the weaker male.
Best course I ever took :)
I had not heard of the less versus fewer distinction mentioned above, and so checked the dictionary; here is the usage note:
Related to your point about Super Bowl commercials, here's an article saying that good targeting makes online ads less effective because targeting removes the need for expensive signaling. Though it doesn't cite any hard data.
This is a very readable and interesting guide, and it may have been dead-on accurate in 1983 when it was written. But the kind of class system he describes, one defined by social signals and not by (say) brute force or even money, can only exist in a unified culture, in which everybody speaks th... (read more)
At least you didn't confuse "discreet" and "discrete". That drives me up the wall!
So, what are you signalling with this post?
Oh, the post was just a means to an end; it's the karma I'll use for signaling.
shminux is obviously signalling "I see what you did there", to elevate his social status.
You're signalling a more-obvious-still "I see what you did there", which presumably isn't worth as much status, but has the added benefit of "calling shminux out", reducing his status gains. Based on karma, your gain seems to be marginally better :)
I'm signalling a tendency to state the obvious verbosely, which, judging from my comment history, isn't worth diddly for status. I'm trying not to speculate on what it means that I make these posts despite that >.>
This is a good, basic, well written introduction to what signaling is. Signaling is a quite vital concept that is often not easy to explain, and so it's nice to have a well written article that explains it well. I definitely will be directing people to this article in the future.
I appreciate the value of the illustrations, but it would be good to find a version that doesn't assume that women are merely materialistic, assessing prospective mates solely on their net worth. Geek communities are often not friendly places for women - some readers will accept the assumptions for the sake of argument, but some are likely to take offence.
Edited when MixedNuts pointed out how confused the original version was. My apologies.
Your numerical model at the beginning looks incoherent.
If employers could never distinguish between smart and stupid, even after hiring, then they could select randomly and pay $70,000 per year (the expected value per hire given the 50:50 ratio). If they can... (read more)
I'm curious, was the Art History comment a dig at Michael Lewis?
What are the costs associated with flowers?
I don't know enough about gardening to have a reasonable opinion on that, but here are some possibilities:
It takes resources in the form of studying fashion, or hiring someone else to do it for you, in order to know which flowers are in, and they're not the ones you expect. Compare to fashion in clothing; I'm probably unfashionable, and this correctly signals that I don't have enough time to keep up with trends or enough hip friends to advise me on them.
The high-status flowers are harder/more expensive to grow than the low status flowers. Compare to his discussion on gravel driveways being higher-status than concrete because they require more maintenance.
The high-status flowers are considered ugly to most untrained people, so they're a net loss unless you know you're associating with people who have been trained in taste. Compare to modern academic music, which will sound unpleasant to an untrained ear; therefore if you like it it signals not only that you yourself are trained, but that you expect the people who judge you based on your musical tastes to be trained. Hence the "any flower too vividly red is prole" comment.
The high-status flowers would look pretentious to low-status people (I realize this is partly an explanatory regress). For example, anyone can give their kid a name that sounds upper-class, but if the kid is poor they risk looking stupid.
There's no game theory in that calculation.
Link to Paul Fussell's “Guide Through The American Class System” essay is dead, here is an archive of the page.
That article was an interesting read but now I'm having a bout of status anxiety :/.
On the other hand:
One of my favorite James quotes:
William James, The Principles of Psychology 1890, Chapter IV
This sentence is a HUGE RED FLAG: “it shattered my illusion that I mostly avoid thinking about class signals, and instead convinced me that pretty much everything I do from waking up in the morning to going to bed at night is a class signal.”
If signaling can explain everything, then it is in the same category as Freudian psychoanalysis—unfalsifiable and therefore useless.
The idea that signaling explains everything leads to the idea that “people who say that they don’t bother with signaling and don’t use the symbols available to them are REALLY just signali... (read more)
Signals by Brian Skyrms is a great book in this area. It shows how signalling can evolve in even quite simple set-ups.
Is "Guide Through the American Class System" supposed to be on the MIRI research guide?
Or where can I find this mailing list to which Yvain refers?
The cost of conveying information should not be described as a "waste".
Does this means all fashion is signalling ?
What? Yes, of course. It is close to the epitome of pure signalling!
If you enjoyed Paul Fussell I highly recommed Bourdieu's "Distinction".
It addresses many of the criticisms about class and taste claims expressed in the comments: It is extremely data-driven instead of relying on 'common knowledge'. It describes not only how different classes' tastes differ but how those difference evolve over time, and not only a single hierarchy of classes but the different fields in which classes operate.
Bourdieu doesn't seem to be part of the lesswrong canon, but could well be. He has an undeserved reputation as (only) a theo... (read more)
One non-obvious implication of signaling theory is how attempts at egalitarian norms/policy can have bad effects. The direct effect of these policies and/or norms is to make it easier for people not possessing some desired characteristic to emulate the behavior of those that do. The effect is to make it harder to signal, and thus results in more resources wasted on signaling games.
More about gardens and signalling: Michael Pollan's Second Nature has a section about class and regional signals in seed catalogues.
This post, by its contents and tone, seems to really emphasize the downside of signaling. So let me play the other side.
Enabling signaling can add or subtract a huge amount of value from what would happen without signaling. You can tweak your initial example to get a "rat race" outcome where everyone, including the stupid people, sends a costly signal that ends up being completely uninformative (since everyone sends it). But you can also make it prohibitively mentally painful for stupid people to go to college, versus neutral or even enjoyable... (read more)
I didn't, and I suspect the percentage of readers who did is much smaller.
Upvote this comment if you didn't notice, or weren't bothered by, the use of “less” instead of “fewer”.
I am also a really fast speed-reader, and my mind tends to ignore these sorts of words.
I remember one time I opened a fortune cookie, glanced at the slip, laughed, and repeated it to my friend, with perfect grammar. He grabbed the slip, actually read the slip (which apparently had had horrible grammar mistakes), and demanded to know how I had managed to read the slip while unintentionally correcting all the grammar mistakes in it.
Upvote this comment if you were jarred by the use of “less” instead of “fewer”.