Defecting by Accident—A Flaw Common to Analytical People

Re­lated to: Ra­tion­al­ists Should Win, Why Our Kind Can’t Co­op­er­ate, Can Hu­man­ism Match Reli­gion’s Out­put?, Hu­mans Are Not Au­to­mat­i­cally Strate­gic, Paul Gra­ham’s “Why Nerds Are Un­pop­u­lar

The “Pri­soner’s Dilemma” refers to a game the­ory prob­lem de­vel­oped in the 1950′s. Two pris­on­ers are taken and in­ter­ro­gated sep­a­rately. If ei­ther of them con­fesses and be­trays the other per­son—“defect­ing”—they’ll re­ceive a re­duced sen­tence, and their part­ner will get a greater sen­tence. How­ever, if both defect, then they’ll both re­ceive higher sen­tences than if nei­ther of them con­fessed.

This brings the pris­oner to a strange prob­lem. The best solu­tion in­di­vi­d­u­ally is to defect. But if both take the in­di­vi­d­u­ally best solu­tion, then they’ll be worst off over­all. This has wide rang­ing im­pli­ca­tions for in­ter­na­tional re­la­tions, ne­go­ti­a­tion, poli­tics, and many other fields.

Mem­bers of LessWrong are in­cred­ibly smart peo­ple who tend to like game the­ory, and de­bate and ex­plore and try to un­der­stand prob­lems like this. But, does know­ing game the­ory ac­tu­ally make you more effec­tive in real life?

I think the an­swer is yes, with a caveat—you need the ba­sic so­cial skills to im­ple­ment your game the­ory solu­tion. The worst-case sce­nario in an in­ter­ro­ga­tion would be to “defect by ac­ci­dent”—mean­ing that you’d just blurt out some­thing stupidly be­cause you didn’t think it through be­fore speak­ing. This might re­sult in you and your part­ner both re­ceiv­ing higher sen­tences… a very bad situ­a­tion. Game the­ory doesn’t take over un­til ba­sic skill con­di­tions are met, so that you could ac­tu­ally ex­e­cute any plan you come up with.

The Pur­pose of This Post: I think many smart peo­ple “defect” by ac­ci­dent. I don’t mean in se­ri­ous situ­a­tions like a po­lice in­ves­ti­ga­tion. I mean in ca­sual, ev­ery­day situ­a­tions, where they tweak and up­set peo­ple around them by ac­ci­dent, due to a lack of re­flec­tion of de­sired out­comes.

Ra­tion­al­ists should win. Defect­ing by ac­ci­dent fre­quently re­sults in los­ing. Let’s ex­am­ine this phe­nomenon, and ideally work to im­prove it.

Con­tents Of This Post

  • I’ll define “defect­ing by ac­ci­dent.”

  • I’ll ex­plain a com­mon out­come of defect­ing by ac­ci­dent.

  • I’ll give some re­cent, mild ex­am­ples of ac­ci­den­tal defec­tions.

  • I’ll give ex­am­ples of how to turn ac­ci­den­tal defec­tions into co­op­er­a­tion.

  • I’ll give some ex­am­ples of how this can make you more suc­cess­ful at your goals.

  • I’ll list some books I recom­mend if you de­cide to learn more on the topic.

Back­ground—On An­a­lyt­i­cal Skills and Rhetoric

From Paul Gra­ham’s “Why Nerds Are Un­pop­u­lar” -

I know a lot of peo­ple who were nerds in school, and they all tell the same story: there is a strong cor­re­la­tion be­tween be­ing smart and be­ing a nerd, and an even stronger in­verse cor­re­la­tion be­tween be­ing a nerd and be­ing pop­u­lar. Be­ing smart seems to make you un­pop­u­lar.
The key to this mys­tery is to rephrase the ques­tion slightly. Why don’t smart kids make them­selves pop­u­lar? If they’re so smart, why don’t they figure out how pop­u­lar­ity works and beat the sys­tem, just as they do for stan­dard­ized tests?
So if in­tel­li­gence in it­self is not a fac­tor in pop­u­lar­ity, why are smart kids so con­sis­tently un­pop­u­lar? The an­swer, I think, is that they don’t re­ally want to be pop­u­lar.
If some­one had told me that at the time, I would have laughed at him. Be­ing un­pop­u­lar in school makes kids mis­er­able, some of them so mis­er­able that they com­mit suicide. Tel­ling me that I didn’t want to be pop­u­lar would have seemed like tel­ling some­one dy­ing of thirst in a desert that he didn’t want a glass of wa­ter. Of course I wanted to be pop­u­lar.
But in fact I didn’t, not enough. There was some­thing else I wanted more: to be smart.

I be­lieve that “defect­ing by ac­ci­dent” is a re­sult of not learn­ing how differ­ent phras­ing of words and lan­guage can dra­mat­i­cally effect how well your point is taken. It’s been a gen­eral ob­ser­va­tion of mine that a lot of peo­ple in highly in­tel­lec­tual dis­ci­plines like math­e­mat­ics, physics, robotics, en­g­ineer­ing, and com­puter sci­ence/​pro­gram­ming look down on so­cial skills.

Of course, they wouldn’t phrase it that way. They’d say they don’t have time for it—they don’t have time for gos­sip, or poli­tics, or sug­ar­coat­ing. They might say, “I’m a re­al­ist” or “I say it like it is.”

I be­lieve this is a re­sult of not re­al­iz­ing how big the differ­ence in your effec­tive­ness will be de­pend­ing on how you phrase things, in what or­der, how well you ap­peal to an­other per­son’s emo­tions. Peo­ple in highly an­a­lyt­i­cal dis­ci­plines of­ten care about “just the facts”—but, let’s face it, we highly an­a­lyt­i­cal peo­ple are a great minor­ity of the pop­u­la­tion.

Sooner or later, you’re go­ing to have some­thing you care about and you’re go­ing to need to per­suade some­one who is not highly an­a­lyt­i­cal. At that point, you run some se­ri­ous risks of failure if you don’t un­der­stand ba­sic so­cial skills.

Now, most peo­ple would claim that they have ba­sic so­cial skills. But I’m not sure this is borne out by ob­ser­va­tion. This used to be a very key part of any ed­u­cated per­son’s stud­ies: rhetoric. From Wik­ied­pia: “Rhetoric is the art of us­ing lan­guage to com­mu­ni­cate effec­tively and per­sua­sively. … From an­cient Greece to the late 19th Cen­tury, it was a cen­tral part of Western ed­u­ca­tion, filling the need to train pub­lic speak­ers and writ­ers to move au­di­ences to ac­tion with ar­gu­ments.”

Rhetoric is now fre­quently looked down upon by highly in­tel­li­gent and an­a­lyt­i­cal peo­ple. Like Paul Gra­ham says, it’s not that in­tel­lec­tu­als can’t learn it. It’s that they think it’s not a good use of their time, that they’d rather be smart in­stead.

Defect­ing by Accident

Thus, you see highly in­tel­li­gent peo­ple do what I now term “defect­ing by ac­ci­dent”—mean­ing, in the pro­cess of try­ing to have a dis­cus­sion, they in­sult, be­lit­tle, or offend their con­ver­sa­tional part­ner. They com­mit ob­vi­ous, blatant so­cial faux pases, not as a con­scious de­ci­sion of the trade­offs, but by ac­ci­dent be­cause they don’t know bet­ter.

Some­times defect­ing is the right course of ac­tion. Some­times you need to break from who­ever you’re ne­go­ti­at­ing with, in­sist that things are done your way, even at their ex­pense, and take the con­se­quences that may arise from that.

But it’s rarely some­thing you should do by ac­ci­dent.

I’ll give spe­cific, clear ex­am­ples in a mo­ment, but be­fore I do so, let’s look at a gen­eral ex­am­ple of how this can hap­pen.

If you’re at a meet­ing and some­one gives a pre­sen­ta­tion and asks if any­one has ques­tions, and you ask point-blank, “But we don’t have the bud­get or skills to do that, how would we over­come that?”—then, that seems like a highly rea­son­able ques­tion. It’s prob­a­bly very in­tel­li­gent.

What nor­mal peo­ple would con­sider, though, is how this af­fects the per­cep­tion of ev­ery­one in the room. To put it bluntly—it makes the pre­sen­ter look very bad.

That’s okay, if you de­cide that that’s an ac­cept­able part of what you’re do­ing. But you now have some­one who is likely to ac­tively work to un­der­mine you go­ing for­wards. A minor en­emy. Just be­cause you asked a ques­tion ca­su­ally with­out think­ing about it.

In­ter­est­ingly, there’s about a thou­sand ways you could be diplo­matic and tact­ful to ad­dress the key is­sue you have—bud­get­ing/​staffing—with­out em­bar­rass­ing the pre­sen­ter. You could take them aside quietly later and ex­press your con­cern. You could phrase it as, “This seems like an amaz­ing idea and a great pre­sen­ta­tion. I won­der how we could se­cure the bud­get­ing and get the team for it, be­cause it seems like it’d be a prof­itable if we do, and it’d be a shame to miss this op­por­tu­nity.”

Just by phras­ing it that way, you make the pre­sen­ter look good even if the op­tion can’t be funded or staffed. In­stead of ex­press­ing your con­cern as a hole in their pre­sen­ta­tion, you ex­press it as a challenge to be over­come by ev­ery­one in the room. In­stead of your un­der­ly­ing point com­ing across as “your idea is un­fea­si­ble,” it comes across as, “You’ve brought this good idea to us, and I hope we’re smart enough to make it work.”

If the real goal is just to make sure bud­get­ing and fund­ing is taken care of, there’s many ways to do that with­out em­bar­rass­ing and mak­ing an en­emy out of the pre­sen­ter.

Defect­ing by ac­ci­dent is lack­ing the aware­ness, tact, and skill to re­al­ize what the sec­ondary effects of your ac­tions are and act ac­cord­ingly to win.

This is a rel­a­tively ba­sic prob­lem that the ma­jor­ity of “nor­mal” peo­ple un­der­stand, at least on a sub­con­scious level. Most peo­ple re­al­ize that you can’t just show up a pre­sen­ter and make them look bad. Or at least, you should ex­pect them to be hos­tile to you if you do. But many in­tel­li­gent peo­ple say, “What the hell is his prob­lem? I just asked a ques­tion.”

This is due to a lack of un­der­stand­ing of so­cial skills, diplo­macy, tact, and yes, per­haps “poli­tics”—which are un­for­tu­nately a re­al­ity of the world. And again, ra­tio­nal­ists should win. If your ac­tions are lead­ing to hos­tility and defec­tion against you, then you need to con­sider if your ac­tions are the best pos­si­ble.

“Why Our Kind Can’t Co­op­er­ate”

Eliezer’s “Why Our Kind Can’t Co­op­er­ate” is a mas­ter­piece. I’m only go­ing to ex­cerpt three parts, but I’d recom­mend the whole ar­ti­cle.

From when I was still forced to at­tend, I re­mem­ber our syn­a­gogue’s an­nual fundrais­ing ap­peal. It was a sim­ple enough for­mat, if I re­call cor­rectly. The rabbi and the trea­surer talked about the shul’s ex­penses and how vi­tal this an­nual fundraise was, and then the syn­a­gogue’s mem­bers called out their pledges from their seats.

Straight­for­ward, yes?
Let me tell you about a differ­ent an­nual fundrais­ing ap­peal. One that I ran, in fact; dur­ing the early years of a non­profit or­ga­ni­za­tion that may not be named. One differ­ence was that the ap­peal was con­ducted over the In­ter­net. And an­other differ­ence was that the au­di­ence was largely drawn from the athe­ist/​liber­tar­ian/​technophile/​sf-fan/​early-adopter/​pro­gram­mer/​etc crowd. (To point in the rough di­rec­tion of an em­piri­cal cluster in per­son­space. If you un­der­stood the phrase “em­piri­cal cluster in per­son­space” then you know who I’m talk­ing about.)
I crafted the fundrais­ing ap­peal with care. By my na­ture I’m too proud to ask other peo­ple for help; but I’ve got­ten over around 60% of that re­luc­tance over the years. The non­profit needed money and was grow­ing too slowly, so I put some force and po­etry into that year’s an­nual ap­peal. I sent it out to sev­eral mailing lists that cov­ered most of our po­ten­tial sup­port base.
And al­most im­me­di­ately, peo­ple started post­ing to the mailing lists about why they weren’t go­ing to donate. Some of them raised ba­sic ques­tions about the non­profit’s philos­o­phy and mis­sion. Others talked about their brilli­ant ideas for all the other sources that the non­profit could get fund­ing from, in­stead of them. (They didn’t vol­un­teer to con­tact any of those sources them­selves, they just had ideas for how we could do it.)
Now you might say, “Well, maybe your mis­sion and philos­o­phy did have ba­sic prob­lems—you wouldn’t want to­cen­sor that dis­cus­sion, would you?”
Hold on to that thought.
Be­cause peo­ple were donat­ing. We started get­ting dona­tions right away, via Pay­pal. We even got con­grat­u­la­tory notes say­ing how the ap­peal had fi­nally got­ten them to start mov­ing. A dona­tion of $111.11 was ac­com­panied by a mes­sage say­ing, “I de­cided to give **** a lit­tle bit more. One more hun­dred, one more ten, one more sin­gle, one more dime, and one more penny. All may not be for one, but this one is try­ing to be for all.”
But none of those donors posted their agree­ment to the mailing list. Not one.

So far as any of those donors knew, they were alone. And when they tuned in the next day, they dis­cov­ered not thanks, but ar­gu­ments for why they shouldn’t have donated. The crit­i­cisms, the jus­tifi­ca­tions for not donat­ing—only those were dis­played proudly in the open.
As though the trea­surer had finished his an­nual ap­peal, and ev­ery­one not mak­ing a pledge had proudly stood up to call out jus­tifi­ca­tions for re­fus­ing; while those mak­ing pledges whispered them quietly, so that no one could hear.

In­deed, that’s a prob­lem. Eliezer con­tinues:

“It is dan­ger­ous to be half a ra­tio­nal­ist.”

And fi­nally, this point, which is mag­nifi­cent -

Our cul­ture puts all the em­pha­sis on heroic dis­agree­ment and heroic defi­ance, and none on heroic agree­ment or heroic group con­sen­sus. We sig­nal our su­pe­rior in­tel­li­gence and our mem­ber­ship in the non­con­formist com­mu­nity by in­vent­ing clever ob­jec­tions to oth­ers’ ar­gu­ments. Per­haps that is why the athe­ist/​liber­tar­ian/​technophile/​sf-fan/​Sili­con-Valley/​pro­gram­mer/​early-adopter crowd stays marginal­ized, los­ing bat­tles with less non­con­formist fac­tions in larger so­ciety. No, we’re not los­ing be­cause we’re so su­pe­rior, we’re los­ing be­cause our ex­clu­sively in­di­vi­d­u­al­ist tra­di­tions sab­o­tage our abil­ity to co­op­er­ate.

On Be­ing Pedan­tic, Sar­cas­tic, Disagree­able, Non-Com­pli­men­tary, and Other­wise Defect­ing by Accident

You might not re­al­ize it, but in al­most all of hu­man civ­i­liza­tion it’s con­sid­ered in­sult­ing to just point out some­thing wrong some­one is do­ing with­out any pref­ace, soft­en­ing, or mak­ing it clear why you’re do­ing it.

It’s taken for granted in some blunt, “say it like it is” com­mu­ni­ties, but it’s usu­ally taken as a per­sonal at­tack and a sign of an­i­mos­ity in, oh, 90%+ of the rest of civ­i­liza­tion.

In these so-called “nor­mal peo­ple’s so­cieties,” cor­rect­ing them in front of their peers will be per­ceived as try­ing to lower them and make them look stupid. Thus, they’ll likely want to re­tal­i­ate against you, or at least not co­op­er­ate with you.

Now, there’s a time and place to do this any­ways. Some­times there’s an emer­gency, and you don’t have time to take care of peo­ple’s feel­ings, and just need to get some­thing done. But sur­fing the in­ter­net is not that time.

I’m go­ing to take some ex­am­ple replies from a re­cent post I made to illus­trate this. There’s always a risk in do­ing this of not be­ing ob­jec­tive, but I think it’s worth it be­cause (1) I tend to read ev­ery re­ply to me and care­fully re­flect on it for a mo­ment, (2) I un­der­stand ex­actly my first re­ac­tions to these com­ments, and (3) I won’t have to re­hash crit­i­cisms of an­other per­son. Take a grain of salt with you since I’m look­ing at replies to my­self origi­nally, but I think I can give you some good ex­am­ples.

The first thing I want to do is take a sec­ond to men­tion that al­most ev­ery­one in the en­tire world gets emo­tion­ally in­vested in things they cre­ate, and are also a lit­tle in­se­cure about their cre­ations. It’s ex­traor­di­nar­ily rare that peo­ple don’t care what oth­ers’ think of their writ­ing, sci­ence, or art.

Crit­i­cism has good and bad points. Great crit­ics are rare, but they ac­tu­ally make works of cre­ation even in cri­tique. A great critic can give back­ground, con­text, and high­light a num­ber of rele­vant main­stream and ob­scure works through his­tory that the piece they’re cri­tiquing re­minds them of.

Good cri­tique is an art of cre­ation in and of it­self. But bad cri­tique—just blind “that’s wrong” with­out ex­plain­ing why—tends to be con­strued as a hos­tile ac­tion and not ac­com­plish much, other than sig­nal­ling that “heroic dis­agree­ment” that Eliezer talks about.

I re­cently wrote a post ti­tled, “Nahh, that wouldn’t work”. I thought about it for around a week, then it took me about two hours to think it through, draw up key ex­am­ples on pa­per, choose the most suit­able, edit, and post it. It was gen­er­ally well-re­ceived here on LW and on my blog.

I’ll show you three com­ments on there, and how I be­lieve they could be sub­tly tweaked.


> I wiz­ened up,
I don’t think that’s the word you want to use, un­less you’re talk­ing about how you fi­nally lost those 20 pounds by not drink­ing any­more.


FWIW, I think posts like this are more valuable the more they in­clude real-world ex­am­ples; it’s kind of odd to read a post which says I had the­ory A of the world but now I hold the­ory B, with­out read­ing about the ac­tual ob­ser­va­tions. It would be like read­ing a his­tory of quan­tum me­chan­ics or rel­a­tivity with all men­tions of things like the laser or dou­ble-slit ex­per­i­ment or Ed­ding or Michel­son-Mor­ley re­moved.


An in­ter­est­ing start, but I would rather see this in Dis­cus­sion—it’s not fully adapted yet, I think...

Now, I spend a lot of time around an­a­lyt­i­cal peo­ple, so I take no offense at this. But I be­lieve these are good ex­am­ples of what I’d call “ac­ci­den­tal defec­tion”—this is the kind of thing that pro­duces a nega­tive re­ac­tion in the per­son you’re talk­ing to, per­haps with­out you even notic­ing.

#1 is kind of clever point­ing out a spel­ling er­ror. But you have to re­al­ize, in nor­mal so­ciety that’s go­ing to up­set and make hos­tile the per­son you’re ad­dress­ing. Whether you mean to or not, it comes across as, “I’m demon­strat­ing that I’m more clever than you.”

There’s a few ways it could be done differ­ently. For in­stance, an email that says, “Hey Se­bas­tian, I wanted to give you a heads up. I saw your re­cent post, but you spel­led “wisen” as “wizen”—easy spel­ling er­ror to make, since they’re un­com­monly used words, but I thought you should know. “Wizen” means for things to dry up and lose wa­ter. Cheers and best wishes.”

That would point out the er­ror (if that’s the main goal), and also en­gen­der a feel­ing of grat­i­tude in who­ever re­ceived it (me, in this case). Then I would have writ­ten back, “Hey, thanks… I don’t worry about spel­ling too much, but yeah that one’s em­bar­rass­ing, I’ll fix it. Much ap­pre­ci­ated. Any­ways, what are you work­ing on? How can I help?”

I know that’s how I’d have writ­ten back, be­cause that’s how I gen­er­ally write back to some­one who tries to help me out. Mu­tual good­will, it’s a vir­tu­ous cy­cle.

Just point­ing out some­one is wrong in a clever way usu­ally en­gen­ders bad will and makes them dis­like you. The thing is, I know that’s not the in­ten­tion of any­one here—hence, “defect­ing by ac­ci­dent.” An­a­lyt­i­cal peo­ple of­ten don’t even re­al­ize they’re show­ing some­one up when they do it.

I’m not par­tic­u­larly both­ered. I get the in­tent be­hind it. But nor­mal peo­ple are go­ing to be ul­tra-hos­tile if you do it to them. There’s other ways, if you feel the need to point it out pub­li­cly. You could “soften” it by prais­ing first—“Hey, some in­ter­est­ing points in this one… I’ve thought about a similar bias of not con­sid­er­ing out­comes if I don’t like what it’d mean by the world. By the way, you prob­a­bly didn’t mean wizen there...”—or even just say­ing, “I think you meant ‘wisen’ in­stead of ‘wizen’”—with links to the dic­tio­nary, maybe. Any of those would go over bet­ter with the origi­nal au­thor/​pre­sen­ter whom you’re point­ing out the er­ror to.

Let’s look at point #2. “FWIW, I think posts like this are more valuable the more they in­clude real-world ex­am­ples; it’s kind of odd to read a post which says I had the­ory A of the world but now I hold the­ory B, with­out read­ing about the ac­tual ob­ser­va­tions.”

This is some­thing which makes peo­ple try­ing to help or cre­ate shake their head. See, it’s po­ten­tially a good point. But af­ter some­one takes some time to cre­ate some­thing and give it away for free, then hear­ing, “Your work would be more valuable if you did (xyz) in­stead. Your way is kind of odd.”

Peo­ple gen­er­ally don’t like that.

Again, it’s triv­ially easy to write that differ­ently. Some­thing like, “Thanks for the post. I was won­der­ing, you men­tioned (claim X), but I won­der if you have any ex­am­ples of claim X so I can un­der­stand it bet­ter?”

That one has—grat­i­tude, no un­nec­es­sary crit­i­cism, ex­plains your mo­ti­va­tion. All of which are good so­cial skill points, es­pe­cially the last one as writ­ten about in Cial­dini’s “In­fluence”—give a rea­son why.

#3 - “An in­ter­est­ing start, but I would rather see this in Dis­cus­sion—it’s not fully adapted yet, I think...”

Okay. Why?

The differ­ence be­tween com­plain­ing and con­struc­tive work is look­ing for solu­tions. So, “There’s some good stuff in here, but I think we could adapt it more. One thing I was think­ing is (main point).”

Be­com­ing More Self-Aware and Strate­gic; Some Prac­ti­cal So­cial Guidelines

From Anna Sala­mon’s “Hu­mans Are Not Au­to­mat­i­cally Strate­gic” -

But there are clearly also heuris­tics that would be use­ful to goal-achieve­ment (or that would be part of what it means to “have goals” at all) that we do not au­to­mat­i­cally carry out. We do not au­to­mat­i­cally:
  • (a) Ask our­selves what we’re try­ing to achieve;

  • (b) Ask our­selves how we could tell if we achieved it (“what does it look like to be a good co­me­dian?”) and how we can track progress;

  • (c) Find our­selves strongly, in­trin­si­cally cu­ri­ous about in­for­ma­tion that would help us achieve our goal;

Anna points out that peo­ple don’t au­to­mat­i­cally ask what they’re try­ing to achieve. You don’t, nec­es­sar­ily, ask what you’re try­ing to achieve.

But I would recom­mend you do ask that be­fore speak­ing up so­cially. At least for a while, un­til you’ve got the gen­eral pat­terns figured out.

If you don’t, you run the risk of an­tag­o­niz­ing and mak­ing peo­ple hos­tile to you who would oth­er­wise co­op­er­ate and work with you.

Now, I’ve heard smart peo­ple say, “I don’t have time for that.” This is akin to say­ing, “I don’t have time to achieve what I want to achieve.”

Be­cause it doesn’t take much time, and it makes you much more effec­tive. Ask­ing, “What am I try­ing to achieve here?” goes a long way.

When com­ment­ing on a dis­cus­sion site, who are you writ­ing for? For the au­thor? For the reg­u­lar read­ers? What’s your point in re­ply­ing? If your main point is just to “get to truth and un­der­stand­ing,” then what should your sec­ondary con­sid­er­a­tions be? If there’s a con­flict be­tween the two, would you pre­fer to en­courage the au­thor to write more, or to look clever by point­ing out a pedan­tic point?

I un­der­stand where you’re com­ing from, be­cause I used to come from the same place. I was the kid who ar­gued with teach­ers when they were wrong, not re­al­iz­ing the long term ram­ifi­ca­tions of that. Peo­ple mat­ter, and peo­ple’s feel­ings mat­ter, es­pe­cially if they have sway over your life, but even if they don’t have sway over your life.

To that, here’s some sug­ges­tions I think would make you more effec­tive:

  • Gen­er­ally, be gra­cious and thank­ful. This goes im­mensely far. Things like start­ing a re­ply with, “Thanks for this” or “Thanks for shar­ing these in­sights.”

  • Prais­ing some­one makes it more likely they’ll ac­cept your crit­i­cisms. “I thought your point A was ex­cel­lent, how­ever point B...”

  • If you’re go­ing to dis­agree, sum­ma­rize the per­son’s main ar­gu­ment be­fore­hand—this has a few pos­i­tive effects. First, it forces you make sure you ac­tu­ally un­der­stand. Se­cond, if the au­thor has a differ­ent main point and wasn’t clear, that comes out. Third, it shows some re­spect that you ac­tu­ally took the time to read and un­der­stand the post. So you could write, “I know your main ar­gu­ment is A, but I wanted to ex­plore your minor point X.”

  • If you think some­thing is wrong, give an ex­pla­na­tion of what would be cor­rect and bet­ter. “I en­joyed this post a lot—thanks for that—but one thing that’s tough for me is that all the ex­am­ples are about mar­tial arts, and I don’t re­ally un­der­stand mar­tial arts so much. Maybe next time you could provide some ex­am­ples from other fields? For in­stance, I re­mem­ber read­ing you’re an ac­coun­tant and you write po­etry, maybe some ex­am­ples from there?”

  • If you point out some­thing is wrong, do your best to make the mis­take-maker not feel stupid. This makes them mas­sively ap­pre­ci­ate that. “Hey, you got your math on ex­am­ple X wrong… I think it ac­tu­ally works to 11.7. Any­ways, I only rec­og­nize that be­cause I made that mis­take dozens of times my­self, it’s a com­mon one to make, just wanted to point it out.”

  • Ex­plain why you care about a point. This has a few pos­i­tive effects. First, it lets the au­thor cater a re­ply to ex­actly what you want. Se­cond, you’d be amazed at how many peo­ple as­sume evil in­tent and worst-pos­si­ble mo­tives—it neu­tral­izes that. Third, it forces you to think through how you’d like things to be, which is again good. “Hey man, I re­ally liked this post, but I won­der if you could have split it into pieces and made it a three-parter? I ask be­cause I surf the web from work, and I can only read in 10 minute chunks… longer posts are harder for me to get through, and I like read­ing your writ­ing.”

  • Con­sider cor­rect­ing some­one pri­vately while prais­ing them pub­li­cly. This com­bi­na­tion has been ob­served to en­gen­der loy­alty and good feel­ings through­out his­tory. I re­cently read an ex­am­ple of a samu­rai en­courag­ing lords to do this from the early 1700′s book “Ha­gakure.” It works.

  • Con­sider drop­ping it al­to­gether if it’s not a big deal. This about learn­ing to pri­ori­tize—I had some­one com­ment on my site think­ing mis­tak­enly that The Rich­est Man in Baby­lon and The Great­est Sales­man in the World were by the same au­thor. It wasn’t, but who cares? It makes no differ­ence. It’s not worth point­ing it out—al­most ev­ery­one has an aver­sion to be­ing cor­rected, so only do it if there’s ac­tu­ally tan­gible gain. Other­wise, go do some­thing more im­por­tant and not en­gen­der the po­ten­tial bad will.

Fol­low­ing some of these sim­ple points will make you much more effec­tive so­cially. I feel like a lot of times an­a­lyt­i­cal and in­tel­li­gent peo­ple study re­ally hard, difficult prob­lems, while ig­nor­ing ba­sic con­sid­er­a­tions that have much more im­me­di­ate and larger im­pact.

Fur­ther read­ing:

Edit: Lots of com­ments on this. 130 and count­ing. The most com­mon crit­i­cism seems to be that adding fluff is a waste of time, in­sincere, and re­duces sig­nal:noise ra­tio. I’d en­courage you to ac­tu­ally try it in­stead of just guess­ing—a quick word of thanks or en­courage­ment be­fore crit­i­ciz­ing cre­ates a more friendly, co­op­er­a­tive en­vi­ron­ment and works well. It doesn’t take very long, and it doesn’t de­tract from S:N ra­tio much, if at all.

Don’t just guess here. Try it out for a month. I think you’ll be amazed at how differ­ently peo­ple re­act to you, and the up­take on your sug­ges­tions and feed­back and abil­ity to con­vince and teach peo­ple. Of course, you can con­struct ex­am­ples of go­ing over­board and it be­ing silly. But that’s not re­quired—just try to make ev­ery­thing 10% more gra­cious, and watch how much your effec­tive­ness in­creases.