Friday, June 13, 2014

What I learned this week - evolutionary psychology

I just finished reading Robert Wright's The Moral Animal. It's an introduction to evolutionary psychology, accompanied by an application of the theory to Darwin's life, as well as some exploration of larger implications. I found the introduction to evolutionary psychology quite stimulating, so I've written up some of the main ideas below.


  1. I know very little about evolutionary psychology. My knowledge is mostly from one book published in 1994, so it may well be incomplete, oversimplified, outdated, or just wrong. (Please correct me if you see errors!) This post should be more of an interest-generator than a reference. 
  2. Some of evolutionary psychology's claims concern why we are the way we are. So these claims - much like those about historical fact - are not verifiable; to accept them is just to think they tell the most compelling story. Evolutionary psychology also makes testable claims about the way we are, by inferring from the conditions of evolution in our ancestral environment. These claims are just hypotheses until they are verified by experiments; we should be careful not to jump to the conclusions just because they feel intuitive.
  3. Often, evolutionary psychology says that humans are disposed to mentalities or behaviors that we consider immoral, unfair, etc. This does not mean it advocates these behaviors! In fact, understanding the source of our natural tendencies can help us to stifle those of which we disapprove. Says Peter Singer, "the more you know about your opponent, the better your chances of winning."


  1. My language below is somewhat fluid between past and present tense. Evolutionary psychology's thesis is basically that this fluidity makes sense: Our brains were shaped by evolution somewhat generic hunter-gatherer societies and haven't changed much since - so we can explain how we are now by looking at the conditions favored by the "ancestral environment". 
  2. Sometimes I say "X should do Y" to mean "the theory would expect that X would do Y". This isn't to be confused with the moral "should".
  3. Finally, the claims below are gross generalizations. They are intended to be describe most people, not all.


p. 41 - The cost to reproduction (without raising of children) for males is very low. For females it is much higher. The benefit in terms of gene-passing on is about the same. So males should be willing to have (uncommitted) sex in almost all circumstances, whereas women should only do so in the most propitious circumstances (when the male has good genes or will help raise the children).

p. 61 - Females will almost always take care of their children - which roughly fixes their level of parental investment. So males should mostly seek out female mates on the basis of good genes and ability of the female to raise the child.. Humans are a high MPI (male parental investment) species, but can vary a lot in how much they invest in the children they have with different partners. So females should seek out males that (along with the earlier characteristics) strongly signal commitment to the relationship. They should also get good at distinguishing committed males from pretenders; in response, the males should get good at acting, including via tricking themselves into feeling committed.

p. 71 - The value of cheating on a spouse: For males, it's to spread sperm as widely as possible. For females, it's to be inseminated by someone with really good genes - whether or not they're committed - and get some other male (thinking it's his kid) to help raise it.

p. 87 - As a marriage progresses, the temptation to abandon it should shift towards a male, who has more reproductive potential to lose by staying together. A female's reproductive potential wanes more quickly.

p. 97 - From a purely evolutionary perspective, polygyny is more egalitarian towards females than males. If a monogamous society becomes polygynous (with multiple females per male), the least desirable males end up with no reproductive partner at all, and the most desirable males, with multiple partners. By contrast, females end up with a more reproductively fit partner, and at most lose a fraction of their partner's parental investment.


p. 157 - We can explain altruism between genetically similar organisms by thinking on the level of the gene. Some gene X can propagate itself not only by increasing the offspring of an organism with X, but also by encouraging X-having organisms to make sacrifices for other organisms likely to have X. For instance, consider a ground squirrel with a gene that makes it yell a warning when predators are near; the individual who yells will be killed by the predator, but fewer of its family members - some of whom probably also have this gene - will die. So by sacrificing an individual, the warning-gene gets passed on more.

p. 164 - Whereas humans of common descent share about 1/2 of their genes, ants share 3/4. So we would expect that ants be more altruistic - their self sacrifice is more likely to benefit individuals with similar genes. Indeed we see this. In the extreme case, consider cells within a human body, which have all the same genetic material. The only way for any of it to get passed on is through the sex cells, and so the entire body cooperates so as to get the sex cells to spread its shared DNA.

p. 187 - Why are we altruistic towards people who don't share much of our genetic material? An early theory called group selectionism suggested that tribes with many patriotic, courageous, sympathetic, etc. individuals would be the most likely to prevail over other tribes - and therefore spread their tendencies. However, this theory doesn't make sense when you consider that the more "noble" males would likely produce fewer offspring than the less noble tribesmen who sleep with others' mates; the tribe would never stabilize at "nobility" in the first place. So why are we altruistic outside of our gene pool?

p. 194 - The answer comes from game theory. A non-zero-sum game is one in which the winners win more than the losers lose. By cooperating in repeated non-zero-sum games, all players can end up better off than if they hadn't cooperated.
For example, consider a chimp who has two bananas but needs one. The chimp could share with a friend who needs a banana but has none. In any single instance, he loses some welfare by giving away the second banana, but not as much as the friend gains; there is an overall increase in welfare from sharing. If these chimps get in the habit of sharing (and are equally needy), then they'll both be better off than if they never shared. And they'll out-compete chimps who fail to share!
Non-zero-sum games are prevalent; some other examples are division of labor and sharing of private information. So we probably have whatever genes would help us enact optimal strategies for playing non-zero-sum games. In many non-zero-sum games, the best strategy is TIT FOR TAT; cooperate until your partner defects, and then defect until she starts cooperating again.
Indeed, our emotions are well-calibrated to help us act according to TIT FOR TAT strategies: When someone does a favor for us, we tend to feel gratitude, i.e. a desire to do a favor for them in return. When we are betrayed, we feel indignant - we don't want to cooperate until the defector makes amends. And we feel especially altruistic towards our friends, the people who we trust to cooperate with us.
And so the general altruism we feel towards people (especially trusted friends) outside of our gene pool is called reciprocal altruism - it's about trading favors.

p. 208 - With respect to reciprocal altruism: It's not so important that we have actually helped our friends; it's important that we have helped them. So we should tend to exaggerate and publicly articulate our sacrifices and the ways we've been wronged, while downplaying (or even forgetting!) how others have helped us.

p. 212 - The genes associated with "conscientiousness" have a heritability of between 0.3 and 0.4 - that is, about a third of differences in conscientiousness among (modern) people can be traced to their genes, and the rest comes from environmental influence. One reason why humans should desire to be known by the community as "nice" or "good" is that they'd like to be perceived as reliable reciprocal altruists. Then the weak heritability of these genes would suggest that the value of reciprocal altruism depends a lot on the society. For instance, in large group settings one might be able to get away with more trickery, but in a small town, everyone can remember who is reliable and who isn't. So its more valuable to be conscientious - and thereby obtain a good reputation - in the small town case.

Social structures

p. 240 - Social hierarchies can emerge as a way to avoid conflict: If A and B both know that A would beat B in a fight over a mate, then they are both better off letting A have the mate and avoiding unnecessary conflict. Genes that encourage this sort of against-needless-conflict hierarchy should prevail.

p. 243 - Nature's mechanism for enforcing social hierarchies seems to be neurotransmitters like serotonin. Higher levels of serotonin encourage more gregarious, socially assertive behaviors; extremely low levels of serotonin are associated with low self-esteem and depression. It appears that - when we anticipate that a conflict will result in a win or a loss - our brain adjusts levels of serotonin accordingly and guides us to the appropriate action. (Maybe serious conflicts are less likely to occur in modern social settings than in the ancestral environment; this could make our serotonin levels poorly calibrated, perhaps explaining why it can be so unreasonably demotivating to learn that someone else is better than us at something)

p. 249 - Status is a resource like anything else, and it can be exchanged in non-zero-sum games. Status assistance may be the biggest gain from reciprocal altruism. We may have a hard-coded tendency towards political patronage! Yikes.

p. 257 - It is normal for humans not just to want their friends and coalitions to win, but to feel that it is morally right. We can look at this as a mechanism for status-trading and reciprocal altruism. And also the source of a lot of inter-tribe conflict.

My main takeaways

  1. Our brain can be hard-coded towards certain biases or dishonesties that we don't know about - ignorance of our own dishonestly would help males convince themselves they will be devoted to a female, help game-players convince fellow game-players they won't defect, and help keep cooperative coalitions together. These biases create inconsistencies between people and groups that think they're being honest, and are therefore an obvious source of conflict.
  2. It's not unreasonable to expect some biological differences between males and females. So pursuing more symmetric gender roles isn't just about changing societal rules; it's also about changing thoughts and behaviors within individuals.
  3. The strength of one's reciprocal altruism may depend strongly on one's upbringing. It would seem that the best way to produce altruistic people is to habituate them to being rewarded/punished them based on accurate assessment of their actions.

No comments:

Post a Comment