You can’t signal to rubes

The word ‘sig­nal­ling’ is of­ten used in Less Wrong, and of­ten used wrongly. This post is in­tended to call out our com­mu­nity on its wrongful use, as well as serve as an in­tro­duc­tion to the cor­rect con­cept of sig­nal­ling as con­trast.

“We choose to go to the moon. We choose to go to the moon in this decade and do the other things, not be­cause they are easy, but be­cause they are hard..”

- John F. Kennedy

Why do pea­cocks grow such large, con­spicu­ous tails? Why do peo­ple take de­grees in sub­jects like Philos­o­phy or Clas­sics, de­spite these sub­jects hav­ing no ob­vi­ous prac­ti­cal value? Why do peo­ple take pains to avoid split­ting in­fini­tives, even though ev­ery­one can un­der­stand split in­fini­tives perfectly well?

Th­ese ac­tivi­ties seem com­pletely pointless, costly and difficult. Para­dox­i­cally, it is prob­a­bly this very difficulty that serves to ex­plain why they are done at all. Take the pea­cock’s tail. A pea­cock that has to strug­gle to sur­vive while drag­ging around a con­spicu­ous tail is clearly at a dis­ad­van­tage. But if he can con­tinue to sur­vive, then clearly he must be pretty strong! So the pea­hens may choose to mate with him rather than the pea­cocks with less con­spicu­ous tails, whose sur­vival is thus a less im­pres­sive feat.

As for clas­sics, get­ting a de­gree in clas­sics may be pointless, but it’s also difficult. It re­quires one to read and mem­o­rize vast chunks of text, and to trans­late these texts be­tween Greek, Latin and English pre­cisely. So a per­son who has a de­gree in clas­sics and got a good mark must be a per­son with a good mem­ory who is able to ex­e­cute tasks pre­cisely. Qual­ities ex­tremely use­ful in a civil ser­vant, the oc­cu­pa­tion where many bud­ding clas­si­cists find them­selves. The rule that you mustn’t split in­fini­tives de­rives from Latin where split­ting in­fini­tives was im­pos­si­ble. So a per­son who doesn’t split in­fini­tives is more likely to be a Latin scholar, with the qual­ities of class and in­tel­li­gence that such a thing im­plies.

Even the de­ci­sion to go to the moon might be ex­plained in this way. Carl Sa­gan made the point that a rocket ca­pa­ble of go­ing to the moon is cer­tainly ca­pa­ble of reach­ing Moscow. And it’s clear why Kennedy in the mid­dle of a Cold War would want to demon­strate such a thing.

When we ex­plain a be­havi­our in this way, we say that the be­havi­our is sig­nal­ling. The agent does not perform a task for its own sake, but to show oth­ers that they pos­sess some im­por­tant qual­ity such as strength, a good mem­ory, or mil­i­tary supremacy. The key fea­tures that a be­havi­our must pos­sess for sig­nal­ling to be a good ex­pla­na­tion are as fol­lows.

  1. The be­havi­our seems pointless. This of course, is a mat­ter of per­spec­tive. Pea­cock’s tails are beau­tiful, clas­sics is in­ter­est­ing, and the moon land­ing was a crown­ing mo­ment of awe­some from hu­man­ity. Pea­cock’s tails seem pointless when viewed as strug­gling to sur­vive and re­pro­duce, clas­sics when viewed as a mat­ter of fi­nance, and the moon land­ing when viewed as a mat­ter of mil­i­tary strat­egy. Pri­ori­ties that genes, most peo­ple, and the Amer­i­can gov­ern­ment might be ex­pected to have.

  2. The be­havi­our re­quires a cer­tain qual­ity. Go­ing to the moon re­quires a good rocket, sur­viv­ing with a con­spicu­ous tail re­quires strength and cun­ning, get­ting a good grade in clas­sics re­quires a good mem­ory. This is ac­tu­ally a stric­ter con­di­tion than nec­es­sary. All we need is for the qual­ity to lower the cost of the be­havi­our. In the ab­sence of su­pe­rior rocket tech­nol­ogy, a moon land­ing could be faked, but only by tak­ing ex­traor­di­nary and costly steps. This was ar­gued by Messrs Mitchell and Webb here.

  3. You want oth­ers to be­lieve you have this qual­ity. Pea­cocks want to look strong and sexy, civil ser­vice ap­pli­cants want to look in­tel­li­gent and dili­gent, Amer­ica wants to look in­no­va­tive and pow­er­ful.

  4. Dishon­est sig­nal­ling isn’t worth it. By dishon­est sig­nal­ling, I mean enag­ing in a sig­nal­ling be­havi­our when you don’t pos­sess the qual­ity you wish oth­ers to think you have. A weak pea­cock who grows a con­spicu­ous tale will be eaten by preda­tors. A stupid per­son who tries to get a de­gree in clas­sics will fail his sub­jects. Fak­ing the moon land­ing car­ries the risk that the con­spir­acy will be ex­posed and the Amer­i­can gov­ern­ment would be­come a laugh­ing stock. A good way of re­mem­ber­ing this crite­rion is the slo­gan “You can’t sig­nal to rubes.”


Un­for­tu­nately, not all pro­posed ex­pla­na­tions in­volv­ing the word “sig­nal­ling” take care to es­tab­lish these four prop­er­ties. Our com­mu­nity seems es­pe­cially guilty of this. The main mi­s­un­der­stand­ing is that it uses ‘sig­nal­ling’ merely to de­note be­havi­ours that trick rubes in to think­ing you’re good. This raises the ques­tion of why there are rubes to trick in the first place. Why haven’t more savvy com­peti­tors eaten their lunch? Here is an ex­am­ple of some­one think­ing that you can sig­nal to rubes:

In other words, it’s all about sig­nal­ing, isn’t it? Man­agers will take ac­tions that ac­tively harm the con­tinued progress of the pro­ject if that ac­tion makes them look “de­ci­sive” and “in charge”. I’ve seen this on many pro­jects I’ve been on, and it took me a while to re­al­ize that my man­agers weren’t stupid or ig­no­rant. It’s just that the or­ga­ni­za­tion I was work­ing in put a higher pri­or­ity on pro­cess than on re­sults. My man­agers, there­fore quite ra­tio­nally did things that max­i­mized their ap­par­ent value in the eyes of their bosses, even if it meant that the pro­ject (and, as a re­sult) the en­tire or­ga­ni­za­tion was hurt.

Here the rube is the man­agers bosses, why are they so stupid as to think that mis­man­age­ment is ev­i­dence of su­pe­rior man­age­ment qual­ities? Why haven’t these idiots been sacked? (This prob­a­bly does oc­cur in real life, but I don’t think “sig­nal­ling” is the right term to de­scribe it. I would de­scribe it as “pan­der­ing to the prej­u­dices of idiots”.)

Another com­ment which falls in to the same trap:

Com­pare the skil­led butcher, who, with no wasted move­ments, cuts his meat just where the joints are, and the flashy butcher, whose flour­ishes make for less skilful and effi­cient cut­ting but send a more im­pres­sive sig­nal.
I agree that the flashy butcher could be­came en­gaged in his cut­ting and lose con­scious­ness of the crowd and his im­pres­sion on it with­out de­creas­ing his sig­nal­ling be­havi­our. If he did so, he might be­come more sincere, but his sig­nal­ling be­havi­our would re­main. For sig­nal­ing is not a con­scious ad­di­tion to his art, which might strip away: skill at cut­ting and skill at sig­nal­ling are wo­ven con­fus­edly to­gether in it.

The trou­ble here is that it pos­tu­lates stupid cus­tomers, just like the pre­vi­ous com­ment pos­tu­lated stupid bosses. A much bet­ter test of butcher qual­ity than flash­iness is how good the meat tastes and how much is pro­duced. An in­tel­li­gent cus­tomer can prob­a­bly test this fairly eas­ily, and would not buy meat from the flashy butcher.

Th­ese uses of “sig­nal­ling” at least have the ad­van­tage that they’re ex­pla­na­tions along eco­nomic lines. The differ­ence be­tween sig­nal­ling and pan­der­ing is the in­tel­li­gence of your au­di­ence. What’s worse is that some peo­ple in our com­mu­nity use the word “sig­nal” to mean “show” or “pre­tend”.

An ex­am­ple:

I may have learned that sig­nal­ling low sta­tus—to avoid in­timi­dat­ing out­siders—may be less of a good strat­egy than sig­nal­ling that I know what I’m talk­ing about.

A low sta­tus per­son that knows what they’re talk­ing about? I sup­pose such things are pos­si­ble… Se­ri­ously though, “sig­nal­ling” is be­ing used to mean “trick­ing peo­ple in to think­ing that you are”. Either you know what you’re talk­ing about or you don’t. At least one of the two op­tions given in the quote will re­sult in you try­ing to trick some­one. We’re sig­nal­ling to rubes again.

Worst of all, some peo­ple use “sig­nal­ling” as a ver­sion of ad hominem. “You just say that to sig­nal.” A com­ment to Over­com­ing Bias’s con­tro­ver­sial post “Gen­tle, Silent Rape” reads:

“I’m sad but not at all sur­prised that so many of these com­ments are to the effect of “HOW DARE YOU EVEN THINK ABOUT THIS TOPIC?!!?!?”
I can only imag­ine how much more frus­trat­ing it must be for Pro­fes­sor Han­son have to deal with in­creas­ingly harsh anti-mind cli­mate that en­velopes the Western world. The all-en­com­pass­ing ide­ol­ogy of poli­ti­cal cor­rect­ness is truly a fear­some jug­ger­naut.
I think the sta­tus sig­nal­ing ar­gu­ments are right on the money, and the rest of the com­ments serve as the proof.”

Let’s go through the crite­ria again:

1. The be­havi­our seems pointless.

This clearly doesn’t ap­ply. The be­havi­our is eas­ily ex­pli­ca­ble. Com­ments might be made out of gen­uine dis­agree­ment, or (more cyn­i­cally) to in­timi­date Han­son and oth­ers away from mak­ing ar­gu­ments like these in the fu­ture.

2.The be­havi­our re­quires a cer­tain qual­ity.

The qual­ity pro­posed was “sta­tus”, but out­rage is cheap. Any fool can be out­raged at a blog post men­tion­ing rape. It doesn’t re­quire ex­cep­tional in­tel­li­gence, charisma, wealth, or fem­i­nist cre­den­tials. You could be home­less and leave an out­raged com­ment just by go­ing to a pub­lic library. You don’t even have to read the post.

3. You want oth­ers to be­lieve you have this qual­ity.

Well this seem­ingly ap­plies. Peo­ple do want to be thought of as be­ing against rape, and high sta­tus. The only trou­ble is that many of the com­ments are left anony­mously.

4. Dishon­est sig­nal­ling isn’t worth it.

This does not ap­ply at all. Even a con­victed rapist could leave an out­raged com­ment.

Clear think­ing re­quires mak­ing dis­tinc­tions. Us­ing the word “sig­nal­ling” to mean “pan­der­ing”, “trick­ing peo­ple”, “show­ing”, or “toe­ing the party line” does noth­ing but lead to con­fu­sion and mud­dle. If you’re go­ing to use jar­gon, use it in its pre­cise sense. That’s what is jar­gon is for, com­mu­ni­cat­ing pre­cisely. Next time you feel like us­ing the word “sig­nal­ling”, ask your­self whether the four crite­ria ap­ply. Re­mem­ber: You can’t sig­nal to rubes.