How to learn soft skills

Ac­quiring some skills is mostly about de­liber­ate, ex­plicit in­for­ma­tion trans­fer. For ex­am­ple, one might ex­plic­itly learn the cap­i­tal of Mis­souri, or the num­ber of miles one can drive be­fore need­ing an oil change, or how to use the quadratic for­mula to solve quadratic equa­tions.

For other skills, prac­ti­tion­ers’ skill rests largely on semi-con­scious, non-ex­plicit pat­terns of per­cep­tion and ac­tion. I have in mind here such skills as:

  • Manag­ing your emo­tions and en­ergy lev­els;

  • Build­ing strong re­la­tion­ships;

  • Mak­ing ro­bust plans;

  • Find­ing an­gles of at­tack on a math­e­mat­i­cal prob­lem;

  • Writ­ing per­sua­sively;

  • Think­ing through charged sub­jects with­out bias;

and so on. Ex­perts in these skills will of­ten be un­able to ac­cu­rately and ex­plic­itly de­scribe how to do what they do, but they will be skil­led nonethe­less.

I’d like to share some thoughts on how to learn such “soft skills”.

Use­ful­ness of non-true stimuli

If you read a chem­istry text­book, it makes sense to ask af­ter each sen­tence: “Is this true?”. If the an­swer is “no”, “no”, “no”, for a suffi­cient num­ber of sen­tences, you should prob­a­bly aban­don that book and look for a bet­ter one. Chem­istry text­books are sup­posed to be made out of state­ments you can trust—state­ments you can add to your file of “trusted ex­plicit claims”, in such a fash­ion as to make you bet­ter at chem­istry. When a book fails at this prop­erty, its main value is lost.

Not so, IMO, for soft skills.

You can test ideas in your “in­ner simu­la­tor”

Your “in­ner simu­la­tor” is CFAR’s ver­sion of the dis­tinc­tion be­tween pro­fes­sion and an­ti­ci­pa­tion. Ba­si­cally, your “in­ner simu­la­tor” is the part of you that can play movies for­ward to de­ter­mine what to an­ti­ci­pate: “Do I have time to turn left be­fore that car reaches me?”; “What will she do, if I ap­proach and say ‘hi’?” (that is: what does my in­ner movie-player show as the next scene, when I play it a movie in which I walk up to her and say ‘hi’?).

Your in­ner simu­la­tor is prob­a­bly more ac­cu­rate than your ex­plicit mod­els in do­mains where you have a rich ex­pe­rience base, such as so­cial phe­nom­ena, and day-to-day phys­i­cal phe­nom­ena. It is prob­a­bly worse in ar­eas where you have good book-learn­ing be­hind you (e.g., you may have an ac­cu­rate con­scious model of the by­stan­der effect, but still mis­model this when you an­ti­ci­pate with­out con­scious ad­just­ment). Your in­ner simu­la­tor is also the place where learn­ing *must* land, if it is to af­fect your au­to­matic sys­tem 1 re­sponses (such as the per­cep­tual pat­terns, and the trig­ger-ac­tion habits, that play into many soft skills).

IMO, most “soft skills” books are not try­ing to add ex­plicit state­ments to your store of “trusted ex­plicit/​ver­bal state­ments”. In­stead, they are try­ing to evoke ex­per­i­ments to try out in your in­ner simu­la­tor—bits that you can then keep, or not, ac­cord­ing to whether they feel promis­ing when you imag­ine try­ing them out. Later, you can try the promis­ing bits out in the ac­tual world.

Ex­am­ple: Roommate

To see how this can work, imag­ine you’re in a tricky so­cial situ­a­tion. Per­haps your room­mate, Fred, is eas­ily offended, and also keeps leav­ing the kitchen in a state of to­tal mess. You’ve brain­stormed a num­ber of op­tions for talk­ing to him, but they all seem likely to end badly, and so you find your­self in the self-help sec­tion of the book­store, look­ing for, well, help.

As you browse, you no­tice a lot of ad­vice that you’ve heard be­fore—ad­vice like “imag­ine what things feel like from Fred’s per­spec­tive” and “ex­plain what’s in it for Fred; find a way to ap­peal to his pride and self-in­ter­est”. You could’ve gen­er­ated a lot of this ad­vice your­self. Nev­er­the­less, much of it is ad­vice you hadn’t ac­tu­ally tried, in Fred’s case. You find your­self moved to ac­tu­ally try it as you read—the sto­ries in the books pull you to ac­tu­ally want to see things from Fred’s per­spec­tive, and you be­gin spon­ta­neously pic­tur­ing how he might be feel­ing. You also find differ­ent sen­tence-stems in your head for how you might start the con­ver­sa­tion—sen­tence-stems seeded, in part, from the sto­ries you read in the books—and some of them seem promis­ing.

Some of the books also con­tain state­ments that, as far as you can tell, are out­right non­sense. One sug­gests that the only rea­son you or Fred have any prob­lems is that you weren’t praised enough as chil­dren. You try on that per­spec­tive as well, but it feels yucky and noth­ing new clicks into place, and so you move along to the next part.

In this case, the books are act­ing, not as a source of trusted in­for­ma­tion, but as spur to your own pro­cess of an­ti­ci­pat­ing, per­ceiv­ing, de­siring, and plan­ning—and, in this way, they are use­ful.

Ex­am­ple: Read­ing a good “woo” book

As I read “Bonds that Make Us Free”, I read many ex­plicit state­ments I dis­agreed with (such as state­ments about a Chris­tian God). I took these in as po­etry: I tried to imag­ine the world the au­thor saw him­self in, and to see my­self in the same world, so as to have more ac­cess to the way he was pars­ing hu­man phe­nom­ena.

The book also con­tained many sto­ries, all of which “rang true” in my in­ner simu­la­tor (they matched the world as I an­ti­ci­pated it), but which formed new pat­terns when placed next to each other. I felt my in­tu­itions up­date as I read—I felt the sto­ries take pat­terns I had pre­vi­ously half-seen, and pull them into full con­scious aware­ness.

I left the book with a changed per­cep­tion of how ra­tio­nal­iza­tion pat­terns af­fect close re­la­tion­ships, and with an in­creased abil­ity to sep­a­rate from my ra­tio­nal­iza­tions and see the peo­ple close to me. I left also with some new free­dom from so­cial shame. Th­ese changes did not come about via trust­ing or defer­ring to the au­thor; they came via try­ing on his per­spec­tive, and find­ing that pieces of the pat­tern he was point­ing to “clicked”.

Try it!

One way to try this is to read a book. But, since books’ use­ful­ness is not just about their ac­cu­racy, it can *also* be sur­pris­ingly effec­tive to just write your own book—or, at least, to write 5 min­utes of it. You can think of this as a way of get­ting sys­tem 1 un­stuck from its de­fault pat­tern.[1]

Set a 5-minute timer (I):

Pick a soft skill, X, that you’d like to get bet­ter at. Then, set a 5-minute timer (yes, an ac­tual one — think­ing for “about 5 min­utes” doesn’t work nearly as well). Spend those 5 min­utes ex­plain­ing to your­self, in writ­ing, how to do X. (For ex­am­ple, if you picked “net­work­ing at con­fer­ences”, you might spend the 5 min­utes brain­storm­ing on what the key tricky bits are, and on strate­gies for nav­i­gat­ing them. E.g., “I’m not sure how to start con­ver­sa­tions. So, at the con­fer­ence, I can watch and see what sen­tence-stems other peo­ple use. I can also just start with ‘Hi, my name is Anna’, or with ‘Oh, are you Dr. so-and-so? I loved your pa­per on such-and-such’. That starter sounds promis­ing, ac­tu­ally; I should try to skim ab­stracts and Google at least a few pa­pers be­fore the con­fer­ence….”)

(You might con­sider giv­ing this a try right now.)

Set a 5-minute timer (II):

This ex­er­cise was cre­ated by Zak Vance, and is one of my fa­vorites. It’s se­ri­ously worth 5 min­utes of your life to try this one out. Pick, again, a soft skill, X, that you’d like to get bet­ter at (e.g., “net­work­ing at con­fer­ences”). Now pick a differ­ent skill, Y, that you’re already highly skil­led at (e.g., “pro­gram­ming” — Y can be a soft skill or any other sort). Now, set a 5-minute timer, and spend 5 min­utes ex­plain­ing (in writ­ing, or aloud to a friend) how skill X is re­ally ac­tu­ally just the same as skill Y, in the sense that any­one who is fluent in skill Y already knows all they need to know to be good at X — they just need to ap­ply their Y-skill to X. (Your goal, as you do this, is to cre­ate a very short guide that en­ables any­one who already knows Y to hit the ground run­ning with X.) You can see my ex­am­ple (taken from the seed “net­work­ing at con­fer­ences is re­ally just the same skill as pro­gram­ming”) in foot­note [2].

It’s im­por­tant, of course, not to be­lieve ev­ery­thing you gen­er­ate in such an ex­er­cise — af­ter all, it was se­cretly writ­ten by a *be­gin­ner* in Skill X. But, again, you can use it to brain­storm ideas to try out, and to thereby get your sys­tem 1 in­tu­itive search pat­tern out of any lo­cal op­tima you may be caught in.

Some good “soft skills” books to try read­ing:

Bonds that Make Us Free, by Terry Warn­er
An Open Heart, by the Dalai La­ma
Feel­ing Good To­gether, by David Burn­s
Eat That Frog, by Brian Tra­cy
How to Win Friends and In­fluence Peo­ple, by Dale Carnegie
Self-ther­apy, by Jay Ear­ley
Fo­cus­ing, by Eu­gene Gendlin (the au­dio­book is much bet­ter than the pa­per, IMO)
The Core LW Se­quences, by Eliezer.

You might also just try go­ing to the book­store and lo­cat­ing a new “soft skills” book, by scan­ning through a few and see­ing if any “speak to you”. Books of the rele­vant sort can be found in the busi­ness, self-help, con­ti­nen­tal philos­o­phy, and spiritu­al­ity sec­tions, as well as in sec­tions fo­cused on par­tic­u­lar soft skills such as writ­ing or prob­lem-solv­ing.

As you read:

As you read a soft skills book (or the re­sults of your own 5-minute timers), you may wish to ask about each para­graph:

  • Is it true? Can I use it to up­date my ex­plicit model of Skill X? (Yes, ex­plicit mod­els are still use­ful!)

  • Even if it’s false—is there some­thing near it that is true?

  • If I free as­so­ci­ate from here, do any of my past ex­pe­riences click into a new fo­cus?

  • What feels fruit­ful/​in­ter­est­ing about this idea? About ideas I can free as­so­ci­ate to from here? Does this im­prove my im­plicit space of hy­pothe­ses?

  • As I read this, can I vi­su­al­ize my­self car­ry­ing out soft skill X in a use­fully differ­ent way? Do I find my in­tu­itions chang­ing?

  • How can I climb as thor­oughly as pos­si­ble into the frame, feel­ing, groove, or wor­ld­view that gen­er­ated this book? And once I do that, does any­thing new click into fo­cus?

Do con­sider adding your fa­vorite books, or book-read­ing strate­gies, in the com­ments!

[1] Sys­tem 1 stuck­ness of this sort seems to be al­most ubiquitous. For ex­am­ple, I type and drive quite a bit, but my typ­ing and driv­ing skill are pretty similar to what they were sev­eral years ago; my con­ver­sa­tional skills im­prove more, but they seem to have some of this same “try­ing the same things again and again” fla­vor. In this con­text, even ran­dom noise seems helpful to jump­start learn­ing.

(“Learned blank­ness” seems re­lated here as well; one some­how gets stuck in an on­tol­ogy; the goal of soft skills books is of­ten to help a per­son jump­start out of it.)

[2] The ex­am­ple I gen­er­ated from this seed, in 90 sec­onds:

Net­work­ing at con­fer­ences is re­ally just the same skill as pro­gram­ming. The three pro­gram­mers’ virtues of Laz­i­ness, Im­pa­tience, and Hubris will take you all the way there, as will ba­sic prin­ci­ples such as an­a­lytic think­ing and code reuse.

Re: Laz­i­ness: Many folks who set out to net­work at con­fer­ences work hard, and push them­selves to talk to lots of peo­ple. But in fact, some con­ver­sa­tions are pre­dictably much higher value than oth­ers. Be Lazy: plan a bit dur­ing your plane-ride to the con­fer­ence (or dur­ing a bor­ing talk!), so that you know *who* you’re hop­ing to have *what kind* of con­ver­sa­tions with, and can po­si­tion your­self ac­cord­ingly.

Also, when you ap­proach a new per­son, see it as an op­por­tu­nity to *prac­tice* and *de­bug* your greet­ing, as a re-us­able code mod­ule, in­stead of treat­ing it as a one-off task to sep­a­rately sweat about and make throw-away code for. And, af­ter­ward, re­view it briefly in your mind — and see if you can cre­ate a way to re­fine it.

You might also look for “test cases”, much as you would when de­bug­ging — you might ap­proach peo­ple who you already know by rep­u­ta­tion, or who a friend of yours has se­cretly already ap­proached, and see if your in­ter­ac­tion pat­tern is similar to what you’ve heard. You can use “test cases” of this sort to find out how your re­sults com­pare to oth­ers, and to gain valuable info for de­bug­ging your own rou­tines.

Note that, to show a real ex­am­ple, the text above is the first thing I gen­er­ated when I gave my­self 90 sec­onds to type from that seed—it is a typ­i­cal case, not a se­lected case—and cer­tainly not a vet­ted guide to net­work­ing. Nev­er­the­less, you can per­haps see how this sort of ra­tio­nal­iza­tion(!) pro­cess might be helpful to seed learn­ing.