Team Cohesion and Exclusionary Egalitarianism

I. Team­work: that shouldn’t be too hard, eh?

Sooner or later, most peo­ple who want to make a differ­ence in the world start think­ing about how to have highly effec­tive teams, team­work, and col­lab­o­ra­tion.

On the sur­face, it seems like it should be sim­ple and straight­for­ward — if you and some­one else both be­lieve in the same cause, you’re both com­pe­tent at your in­di­vi­d­ual roles, and you get along well be­tween the two of you, you should be able to be an effec­tive team.

Find­ing a third mem­ber of the team would just mean get­ting some­one else who cares about the cause, is com­pe­tent, and gets along with the two of you. And so on. Build­ing teams should be easy and straight­for­ward, no?


Ah, were it so sim­ple! But any­one who ac­tu­ally at­tempts to do so, finds out that it’s not the case.

I’ve been kick­ing around an idea about effec­tive or­ga­ni­za­tions for a year or two now, draw­ing on both per­sonal ex­pe­rience in team-build­ing and hiring and man­age­ment, as well as his­tor­i­cal case stud­ies.

I was prompted to write this af­ter read­ing the (ex­cel­lent) A Dialogue on Ra­tion­al­ist Ac­tivism

“Well, for one thing, peo­ple would au­to­mat­i­cally pat­tern-match pretty much any at­tempt at form­ing an or­ga­ni­za­tional struc­ture to a “re­li­gion” or a “cult” even though what’s ac­tu­ally be­ing at­tempted is the literal op­po­site of those things. When it comes to ac­tual for­mal doc­u­ments spec­i­fy­ing the ob­jec­tives and struc­ture of the or­ga­ni­za­tion, peo­ple would get end­lessly caught up in rel­a­tively in­con­se­quen­tial choices of lan­guage or fo­cus, per­pet­u­ally bick­er­ing over the last 1% of lin­guis­tic dis­tinc­tion that sep­a­rates their aims. You would think peo­ple who prize ra­tio­nal­ity would be able to shield them­selves from the nar­cis­sism of small differ­ences, but I sus­pect not, in re­al­ity.”

Now, if you’d never ac­tu­ally tried to build a team in the real world to work on a prob­lem, you might think seems in­sane. “...per­pet­u­ally bick­er­ing over the last 1% of lin­guis­tic dis­tinc­tion that sep­a­rates their aims...” — that wouldn’t ac­tu­ally hap­pen, right? What a waste of time, no?

II. Look­ing at the prob­lem from differ­ent angles

And yet, you set out into the world to build a team and get col­lab­o­ra­tion go­ing at scale, and, well, this of­ten hap­pens.

There’s a lot of ways you could an­a­lyze this prob­lem and look to over­come it.

You might look at it as a per­son­nel challenge — do we have the right peo­ple and right mix of cog­ni­tive styles here?

You might look at it as an in­cen­tives challenge — how do we in­cen­tive peo­ple mak­ing tan­gible con­tri­bu­tions and dis­in­cen­tivize, umm, “per­pet­ual bick­er­ing.”

You might look at it as skills challenge — how do we im­prove all team­mates’ com­mu­ni­ca­tion skills, strat­egy, and pri­ori­ti­za­tion?

The “per­pet­ual bick­er­ing over the last 1%” prob­lem is very real, and figur­ing out how to nav­i­gate it is a tricky challenge. Differ­ent fields and differ­ent or­ga­ni­za­tions at­tempt to solve it in differ­ent ways.

In large open-source soft­ware pro­jects, it’s of­ten solved by hav­ing a sin­gle fi­nal de­ci­sion­maker — the half-jok­ing “Benev­olent Dic­ta­tor for Life” ti­tle be­ing a com­mon one in large-scale OSS pro­jects.

III. Ex­clu­sion­ary Egalitarianism

In this post, I’d like to pre­sent one re­cur­ring pat­tern that at times leads to ex­cep­tional team­work, and is rather coun­ter­in­tu­itive — ex­clu­sion­ary egal­i­tar­i­anism.

On a quick look, this phrase seems like an oxy­moron — when we tend of think of egal­i­tar­ian groups, we tend to as­sume they’re in­clu­sive as a de­fault.

But I think these are ac­tu­ally two sep­a­rate spec­trums:

Egal­i­tar­ian <-> Hierarchical

In­clu­sive <-> Exclusive

I would strongly sus­pect that that prefer­ences for egal­i­tar­ian com­mu­ni­ca­tion and struc­tures highly cor­re­lates with prefer­ences to­wards in­clu­sivity, and like­wise, prefer­ences to­wards hi­er­ar­chi­cal com­mu­ni­ca­tion and struc­tures also cor­re­lates with prefer­ences to­wards ex­clu­sivity.

If that’s true, then you would tend to see two types of ap­proaches to or­ga­ni­za­tions dom­i­nat­ing the world — first, an in­clu­sive and egal­i­tar­ian paradigm that has more mem­bers, more dis­cus­sion, and less clearly defined and for­mal lead­er­ship struc­tures; sec­ond, a more ex­clud­ing and hi­er­ar­chi­cal paradigm that has fewer mem­bers, less dis­cus­sion, and a more clearly defined for­mal lead­er­ship struc­ture.

A stereo­typ­i­cal ex­am­ple of the first might be the 1960s-1970s peace move­ments in the United States; a stereo­typ­i­cal ex­am­ple of the sec­ond might be a very tra­di­tional Ja­panese cor­po­ra­tion in a well-es­tab­lished field run­ning on Con­fu­cian norms.

And yet, it seems like there’s a rare and coun­ter­in­tu­itive pat­tern that some­times out­performs tremen­dously — a high stan­dards bar and ex­clu­sivity to­wards join­ing, fol­lowed by very much egal­i­tar­ian norms once peo­ple have made it.

A real world ex­am­ple might help illus­trate.

Take this ac­count of the Delta Force se­lec­tion pro­cess, from Com­mand Sergeant Ma­jor Eric Haney; em­pha­sis added —

“Fif­teen min­utes later, we halted along the edge of what looked like a drop zone. “Hol­land DZ,” I heard some of the Fort Bragg troops say as we dis­mounted.

Sergeant Ma­jor Shu­mate was stand­ing nearby wear­ing khaki pants, a Hawaiian shirt, and a Panama hat. “Fall in on me, ladies!” he called. “Make it six ranks, I want a tight for­ma­tion.

As we were form­ing up, some­body walked up with a cam­era and a tri­pod and pre­pared to take a pho­to­graph of the for­ma­tion.

“What the hell’s this about, Walt?” an anony­mous voice squawked from some­where in the crowd.

“This is go­ing to be the ‘be­fore pic­ture’ of this group, my young dar­lings, and we’ll take a sec­ond shot in a few weeks,” Shu­mate replied.

I was aghast that some­one had called a sergeant ma­jor by his first name, but Shu­mate seemed to take no offense. But when a small cluster of the men started laugh­ing and ya­hoo­ing within the group, like this was some sort of stupid joke, Shu­mate be­came steely. The joc­u­lar­ity in his voice came to a screech­ing halt and he growled in a vol­canic tone, “Yeah, well, we’ll see who’s laugh­ing when you moth­erf***ers are finished, and the next pic­ture I take is about half the size of this front rank stand­ing here. The se­ri­ous ones will be in that photo and the rest of you s***birds will be back home, ly­ing to your team­mates about why you didn’t make it. So since you loud­mouths — and a bunch more of you be­sides — won’t be here for the ‘af­ter photo,’ I’ll just have my laugh at you d***heads right now — ha-f***ing-ha!”

It was a sober group he ad­dressed now as he called us to at­ten­tion. The pho­tog­ra­pher took the shot and de­parted.”

If you’re fa­mil­iar with gen­eral mil­i­tary deco­rum at all, that’s already a sur­real scene — the Sergeant Ma­jor run­ning Delta Force se­lec­tion showed up be­fore a gru­el­ing test in a Hawaiian shirt… and across ranks, peo­ple just re­fer to each other by their first names in­stead of their full ti­tles… this is un­usual.

But ac­tu­ally, you find this pat­tern across nearly all elite Amer­i­can Spe­cial Forces type units — (1) an ex­ceed­ingly difficult bar to get in, fol­lowed by (2) in­cred­ibly loose, in­for­mal, col­le­gial norms with nearly-in­finitely less em­pha­sis on hi­er­ar­chy and bu­reau­cracy com­pared to all other mil­i­tary units.

To even “try out” for a Spe­cial Forces group like Delta Force or the Navy SEAL Teams, you have to be among the most ded­i­cated, most phys­i­cally fit, and most com­pe­tent of sol­dier.

Then, the se­lec­tion pro­ce­dures are in­cred­ibly in­tense — only around 10% of those who at­tend se­lec­tion ac­tu­ally make the cut.

This is, of course, ex­clu­sion­ary.

But then, seem­ingly para­dox­i­cally, these or­ga­ni­za­tions run with far less hi­er­ar­chy, for­mal au­thor­ity, and tra­di­tional mil­i­tary deco­rum than the norm. They run… far more egal­i­tar­ian than other tra­di­tional mil­i­tary unit.

IV. As­sorted Challenges and De­sign Tradeoffs

Go­ing back to the situ­a­tion Mori­d­i­na­mael de­scribed —

“When it comes to ac­tual for­mal doc­u­ments spec­i­fy­ing the ob­jec­tives and struc­ture of the or­ga­ni­za­tion, peo­ple would get end­lessly caught up in rel­a­tively in­con­se­quen­tial choices of lan­guage or fo­cus, per­pet­u­ally bick­er­ing over the last 1% of lin­guis­tic dis­tinc­tion that sep­a­rates their aims.”

If we search out the root causes of this “per­pet­ual bick­er­ing,” we can find a few right away —

*When there’s low stan­dards of trust among a team, peo­ple tend to ad­vo­cate more strongly for their own prefer­ences. There’s less con­fi­dence on an in­di­vi­d­ual level that one’s own goals and prefer­ences will be reached if not strongly ad­vo­cated for.

*Ideas — es­pe­cially new ideas — are no­to­ri­ously difficult to eval­u­ate. When there’s been no ob­jec­tive stan­dard of perfor­mance set and achieved by peo­ple who are work­ing on strat­egy and doc­trine, you don’t know who has the abil­ity to ac­tu­ally im­ple­ment their ideas and see them through to con­clu­sion.

*Gen­er­ally at the idea phase, peo­ple are max­i­mally ex­cited and en­gaged. Peo­ple are of­ten un­able to model them­selves to know how they’ll perform when the en­thu­si­asm wears off.

*In the ab­sence of pre­vi­ously demon­strated com­pe­tence, peo­ple might want to show they’re fit for a lead­er­ship role or key role in de­ci­sion­mak­ing early, and might want to (per­haps sub­con­sciously) demon­strate prowess at mak­ing good ar­gu­ments, ap­pear­ing smart and eru­dite, etc.

And of course, many more is­sues.

Once again, this is of­ten re­solved by hi­er­ar­chy — X per­son is in charge. In the ab­sence of ev­ery­one agree­ing, we’ll do what X says to do. Be­cause it’s bet­ter than the al­ter­na­tive.

But the trade­offs of hi­er­ar­chi­cal or­ga­ni­za­tions are well-known, and hi­er­ar­chi­cal lead­er­ship seems like a fit for some do­mains far moreso than oth­ers.

On the other end of the spec­trum, it’s easy when be­ing egal­i­tar­ian to not ac­tu­ally have de­ci­sions get made and fail to have valuable work get­ting done. For all the flaws of hi­er­ar­chi­cal lead­er­ship, it does tend to re­solve the “per­pet­ual bick­er­ing” prob­lem.

From both per­sonal ex­pe­rience and a pretty deep im­mer­sion into the his­tory of suc­cess­ful or­ga­ni­za­tions, it looks like of­ten an an­swer is an in­cred­ibly high bar to join­ing fol­lowed by largely de­cen­tral­ized, col­lab­o­ra­tive, egal­i­tar­ian de­ci­sion­mak­ing.

Of course, in­clu­sivity-ex­clu­sivity and egal­i­tar­i­anism-hi­er­ar­chy are both spec­trums you can fall at many places along, but I think the fol­low­ing or­ga­ni­za­tions all had some mea­sure of this —

*The Los Alamos Lab­o­ra­tory: The key site for the Man­hat­tan Pro­ject.

*Bell Labs: In­vented a lion’s share of mod­ern com­put­ing.

*Lock­heed Skunk Works: The fastest high-perfor­mance air­craft de­vel­op­ment pro­ject of all time.

*The afore­men­tioned Delta Force and Navy SEAL Teams: Elite mil­i­tary units who reg­u­larly do seem­ingly im­pos­si­ble things.

*Bridge­wa­ter As­so­ci­ates: The largest and most prof­itable hedge fund of all time.

*DARPA: The key or­ga­ni­za­tion in in­vent­ing the In­ter­net and Global Po­si­tion­ing Sys­tem (GPS), among other note­wor­thy achieve­ments.

To go back a lit­tle fur­ther and time, and to pick more “so­cially” fo­cused groups which were very suc­cess­ful, re­gard­less of how you feel about their aims —

*The Je­suit Order: Be­came more hi­er­ar­chi­cal over time, but started sur­pris­ingly egal­i­tar­ian if you read the his­tory.

*Bol­she­viks: Fa­mously split from Men­she­vik fac­tion of RSDLP over in­clu­sive/​ex­clu­sive ar­gu­ments. Took over Rus­sia and then a large por­tion of the Earth. Again, be­came more hi­er­ar­chi­cal over time.

V. Con­clud­ing and Thinking

Ob­vi­ously, both the words and con­cepts around modes of co­op­er­a­tion and ac­tion can be very emo­tion­ally-charged, but if you step back for a mo­ment, you can see how there’s as­sorted de­sign challenges in­volved in build­ing strong teams that are effec­tive at ad­vanc­ing their mis­sion.

As the origi­nal au­thor noted,

“You would think peo­ple who prize ra­tio­nal­ity would be able to shield them­selves from the nar­cis­sism of small differ­ences, but I sus­pect not, in re­al­ity.”

There’s many de­sign trade­offs in build­ing suc­cess­ful teams and or­ga­ni­za­tions — it’s hard to do.

And get­ting past hag­gling over triv­ial­ities doesn’t guaran­tee suc­cess — quite the con­trary, it’s merely table stakes and a pre­cur­sor for ac­tion and achieve­ments to fol­low.

It’s worth pe­ri­od­i­cally think­ing through and re­search­ing, un­der­stand­ing that differ­ent ways of struc­tur­ing an or­ga­ni­za­tion, differ­ent ways of com­mu­ni­cat­ing, and differ­ent ways of se­lect­ing team­mates is pos­si­ble. Em­piri­cally, some meth­ods work bet­ter than oth­ers at build­ing strong teams that go out and do a lot of good in the world.

I have my own prefer­ences, of course, as does ev­ery­one. Per­son­ally, I greatly pre­fer egal­i­tar­ian and open di­a­log, com­mu­ni­ca­tions, and de­ci­sion­mak­ing within a team — but with the pre­req­ui­site that there’s a high de­gree of trust and demon­strated com­pe­tence among all team mem­bers.

For large-scale challenges, in­di­vi­d­ual perfor­mance even­tu­ally falls off in im­por­tance rel­a­tive to ex­cel­lent team dy­nam­ics, and ex­cel­lent team dy­nam­ics are a no­to­ri­ously hard thing to get right.

It’s definitely worth think­ing about care­fully, study­ing his­tory to learn from past suc­cesses, test­ing differ­ent meth­ods, and an­a­lyz­ing what went right and went wrong.

There’s many differ­ent ways to build strong teams, and even more ways to fail to build strong teams. If you’re look­ing to do some­thing on a very large scale, it’s even­tu­ally one of the most im­por­tant things you’ll need to get right.