Robert Sapolsky’s *Behave*

by on May 11, 2017 at 8:48 am in Books, Science | Permalink

The subtitle is The Biology of Humans at Our Best and Worst.  Sapolsky is tenured in biology and neuroscience at Stanford, and winner of a MacArthur genius grant.  This book is a very impressive compendium of what we know about the social sciences, as might be rooted in behavioral biology and related fields.  The topics include violence, altruism, cooperation, gene-environment interactions, and many more topics along the usual lines.  It’s not a “here is my big idea” book, but rather “here is how we think about social phenomena.”  In the conclusion, Sapolsky writes: “If you had to boil this book down to a single phrase, it would be “It’s complicated.”  Nothing seems to cause anything, instead everything just modulates everything else.”  Those are two very good sentences.

This is likely to be one of this year’s major social science books, and many of you should buy it, but it’s flaw is that there’s no particular claim you are forced to come to terms with.

1 Anonymous May 11, 2017 at 9:04 am

Is the new style guide that “we may nest “quotes” within quotes,” without using single-quotes? I sometimes do, but I always feel guilty about it. Copy paste laziness.

I guess I approve of the recursion tho


2 P May 11, 2017 at 9:09 am

Sapolsky is a pinhead. If there are no main effects, relatives will not resemble each other. But they do and the effect is strong across the board and linearly related to the degree of shared genes.


3 Anonymous May 11, 2017 at 9:19 am

You used a very soft word there. “Resemble.”

They resemble, rather than copy, clone or match, because “it’s complicated.”


4 P May 11, 2017 at 9:25 am

No. They resemble each other to varying degrees because, firstly, only MZ twins are clones of each other and, secondly, neither heritability nor broader familiality are ever 100%.


5 Anonymous May 11, 2017 at 9:31 am

There is that word “resemble” again, now even diluted “to varying degrees.”

.. probably because it’s complicated.


6 P May 11, 2017 at 9:54 am

“Resemble to varying degrees” means that phenotypic correlations between individuals are directly proportional to the degree that they share genetic variants with each other. It’s not complicated. It’s very simple.

7 Pshrnk May 11, 2017 at 10:00 am

@ Anonymous Why engage with P? Obviously P is for Pinhead.

8 Anonymous May 11, 2017 at 10:08 am

I guess I hoped that P would see that he is making claims that sound definite, without being. When I was young I “resembled” all the other nerds in the computer room. Shrug.

9 P May 11, 2017 at 10:17 am

All I am saying that human phenotypic variation is strongly influenced by additive genetic effects. This a completely mundane fact replicated any number of times. It also contradicts Sapolsky’s claim which I interpret as a denial of the existence of strong genetic main effects. It’s the modulating effects that are difficult to find, let alone replicate, not the main effects.

10 john May 12, 2017 at 7:59 am

And genetic transfer is part of the questions social science study? So physcal traits are inherrited. What is the implication for our social interactions and the structures of or social institutions? Is it a clearly causal one traced to the genes?


11 D May 11, 2017 at 10:46 am

Sapolsky is a brilliant science communicator when he wants to be, but his old Teaching Company lectures, which were otherwise fantastic, show him to be anti-science when it comes to PC topics. E.g., he strongly dismisses IQ as anything meaningful in those lectures. You can’t expect anything bold out of a mind like that.


12 Darf Ferrara May 11, 2017 at 12:30 pm

A more complete description of Sapolsky’s view of IQ is that it is complicated.

Sapolsky has a more technical version of the course on Biology and Human Behaviour on iTunes for those interested.


13 Anonymous May 11, 2017 at 3:27 pm

Just happened to see:

That heading might overstate it a bit, but the low correlations for things like “income” and “happiness” might surprise some.


14 Per Kurowski May 11, 2017 at 9:34 am

“It’s complicated.” Nothing seems to cause anything, instead everything just modulates everything else.”

This underlines beautifully the unbelievable hubris of regulators when deciding to impose risk weighted capital requirements on banks


15 FG May 11, 2017 at 11:54 am

Seems like a big ol’ jump from “things are complicated” to “we understand nothing”.


16 NeedleFactory May 11, 2017 at 12:10 pm

Not just the hubris, but the impossibility! See Maymin’s Why Financial Regulation is doomed to Fail:


17 prior_test2 May 11, 2017 at 9:36 am

‘but rather “here is how we think about social phenomena.”

One hopes he explores that whole ‘you’re fired!’ social phenomena and its apparent ongoing appeal in governing the U.S.


18 Anonymous May 11, 2017 at 9:43 am

It’s complicated. Like steam catapults. Or maybe the digital.


19 Rimbaud May 11, 2017 at 10:23 am

phenomena is plural, you mean “that phenomenon”


20 Thor May 11, 2017 at 11:50 am

What? You mean Trump is only going to do it once?


21 rayward May 11, 2017 at 10:39 am

Human behavior is complex, which may explain the paradox of preferring simple answers for it. For some reason, Jonathan Haidt comes to mind. Why did Trump fire Comey? It’s complex. When I was in college, long ago, social studies scholars were partial to using boxes to explain behavior, the box a visual representation of the set within it, overlapping boxes a visual representation of the commonality of the two sets within the overlapping boxes, and so on. The Box Method was popular not only in such fields as psychology (they were crazy for boxes) but also in fields such as international relations. I especially liked the boxes that were in color. As I looked on in class at the well-ordered boxes my professor in international relations had drawn, I could imagine life as a diplomat, drawing boxes in preparation for an important meeting with some foreign diplomat, and wondered if the two diplomats shared their boxes and that’s how peace is maintained in the world. Who knew that maintaining peace between adversaries was so simple: the Box Method. Maybe somebody should teach Trump the Box Method for making decisions. He’d probably like the colored boxes. I did when I was in college.


22 Captain Obvious May 11, 2017 at 11:29 am

So this guy received a genius grant and the best he can say is: “If you had to boil this book down to a single phrase, it would be “It’s complicated.”  -> Isnt this the definition of a ZMP worker?


23 FG May 11, 2017 at 11:56 am

Seems somewhat unfair to judge the whole work by the single 5-syllable sentence that’s intentionally reductive.


24 Captain Obvious May 11, 2017 at 1:01 pm

I know, but if this is the best insight he has then …. we are in ZMP research territory no?


25 Anonymous May 11, 2017 at 1:20 pm

Certainly not, if it is the correct answer and refutes those who say “It’s simple!”


26 Todd Kreider May 11, 2017 at 11:47 am

I just watched a youtube clip where Sapolsky tells his students that a huge study of siblings showed that the oldest had the higher I.Q. on average and asks his students why.

After a few minutes of guessing reasons, Sapolsky explains to the class that despite “impeccable science”, that is, it was a 250,000 person study, the difference in I.Q. was only 2.3 points. “Totally reliable (statistics) is different from say, ‘important’.”

But null results or near null results *are* important, even if they usually don’t get published. Sapolsky doesn’t seem to get that.


27 Todd Kreider May 11, 2017 at 12:07 pm

( I should have written: ‘null results can be important’…)


28 Steve Sailer May 11, 2017 at 8:10 pm

2.3 IQ points seem fairly important to me.


29 Todd Kreider May 11, 2017 at 10:53 pm

I agree. With a huge study like that, 2.3 points is quite small but enough that it is interesting. Sapolsky says in the clip that if you sneeze, you lose 2 points. Fine, but it isn’t like the second sibling is always sneezing on the test as the older sibling is not.


30 igorrr May 11, 2017 at 12:44 pm

To say “it’s complicated” tends to make causes sound similar. Let’s make it a betting proposition: Take 200 infants and send them to an unknown country to be raised: 100 are children of high IQ tall parents, 100 are children of low IQ short parents. If we knew nothing about what random conditions they would be raised in we would have to bet that these infants’ lifetime outcomes would correlate strongly with their parents IQ and height. Even stronger, if these children were randomly assigned to a first world country with the caveat that the first group were only adopted by average middle class, median income parents and the second group were only adopted by families with top 10% incomes, we should still bet that the first group will grow up to have higher IQ and height and even higher incomes. Anyone care to bet otherwise? Cf. Sacerdote on Korean adoptees.


31 wiki May 11, 2017 at 12:53 pm

To those who say “It’s complicated,” do we give the same latitude to the discussion of climate change? The models and tests of climate change over a short (less than a century) time period are far, far weaker than the studies on biology, behavior, IQ. The fact that the recent slowdown in warming was not forseen by the consensus should give us pause. It doesn’t deny that warming has happened nor that more warming is likely but it makes it probable that we know little about the parameters affecting near to medium term warming and about the precise contribution of human action to that warming trend. Worse, it means that claims of how changes in policy would control warming in the next century should be taken with more than a grain of salt.

After all, we have a very poor understanding of how to intervene to help low performing children raise their long run performance and IQ in the US (Roland Fryer’s results are mostly negative) yet we should trust climate scientists to make economic and technological policy about warming?


32 Steve Sailer May 11, 2017 at 8:11 pm

“It’s complicated” is my personal excuse for why I don’t pay attention to climate science; but I don’t go around telling other people not to pay attention to the subject.


33 DevOps Dad May 12, 2017 at 11:42 am

While it is complicated with humans to compare genotypes and phenotypes, a simplified, yet convincing test can be made with dogs, and the test could cost around $100,000.

Dog breeders have already done our work for us. The left tail of the IQ normal distribution of the smartest breeds of dog (Border Collie, Poodle) overlaps the right tail of the IQ normal distribution of the dumbest breeds of dog (Bulldog, Afghan). Take 20 just weaned puppies from the smartest and dumbest breeds and provide all with identical complex training and record everything.

You can even provide the dumbest breeds with 50% more training, special foods or a more comfortable environment. When the dogs are two years old, test them on what they were taught.

You can make monetary bets with prior2 or Hazel Meade on what they think the outcome would be.


34 thfmr May 11, 2017 at 1:16 pm

I wonder if academics would start being honest about IQ questions if their fellowships and professorships and chairmanships were taken away en masse and given to people of more favored demographics.

“Maybe if I sit here quietly and play along, they won’t come for mine…”

Canada isn’t f**king around in this regard:


35 Steve Sailer May 11, 2017 at 8:11 pm

“The Biology of Humans at Our Best and Worst”

That sounds problematic!


36 zztop May 11, 2017 at 8:52 pm

What do any of these commenters above know and offer, when compared to the extraordinary intellect and prodigious research output of Dr. Sapolsky? Literally, nothing.


37 Michael May 12, 2017 at 2:59 pm

nobody wants to complain about grammar? TC, I didn’t expect this from you:
“but it’s flaw is that…”


38 g ruqt May 24, 2017 at 4:58 pm

Professor Cowen,

Appreciate the recommendation however you missed a / the observation critical to our times from Dr. Sapolsky. Re-read the chapter on adolescent risk taking and re-read the news briefs on a 22 year old incipient suicide bomber not viewed as a threat. Would you not prefer that intelligence services factor adolescent behavior into their assessment?


Leave a Comment

Previous post:

Next post: