Robert Sapolsky on the Evolution of Behavior

A great way to learn about animal (including human) cognition is to develop a greater understanding of their behaviors from the point of view of different sciences. As I wish to work in an interdisciplinary environment, I’m required to be aware of the different perspectives those fields offer. Stumbling across Robert Sapolsky’s lecture about Behavioral Biology, I felt inclined to take some notes for my future work and decided to share those notes with you.

You can watch the whole lecture here:


On Darwin’s Theory of Evolution


The Logic of the Darwinian Evolution works with four premises:

1. There are traits that are inheritable
2. There is variability in those traits
3. Some versions of those traits are more adaptive than others
4. There is a possibility of mutation

With this in mind, it is important to state that it’s all about reproduction and not about the survival of the individual. The individual’s “goal” or “desire” is to maximize the number of the copies of  its own genes in the next generation. However, speaking in such a way is just a heuristic. There is no intent or plan involved.


How can those principles be applied to behavior?

There are different ways to ensure that the own genes are passed on. Sapolsky explains three possibilities:

  1. Individual selection
  2. Kin selection
  3. Reciprocal altruism

The most direct way to pass on the own genes is quite obvious. The individual itself has to reproduce. The likelihood of being successful in that regard is influenced by the mechanics of natural and sexual selection.

Kin selection comes into play, because we share a certain amount of genes with our relatives. The closer the family relation between the individuals the more genes they share, so there is a mathematic to how this works. J.B.S. Haldane is said to have joked that he would willingly die for two brothers or eight cousins.

Reciprocal altruism describes non-competitive and cooperative behavior towards other non-related individuals and, unsurprisingly, animals engage in it, because it is beneficial to do so. As it is a fairly complex behavior, it requires the individual to have certain properties and abilities: They have to be smart (remember and recognize other individuals and their actions, detect cheating behavior, …), social, and long-lived.



How can an individual optimize its success rate: A short introduction to Game Theory


The Prisoner’s Dilemma and the (forgiving) “Tit for Tat”-strategy:

How is the individual supposed to act to maximize the number of the copies of  its own genes in the next generation? When is it supposed to cheat? When would it be more helpful to cooperate? The Prisoner’s Dilemma (formalized by A. W. Tucker) is one of the more popular examples of game theory and is helpful to illustrate which strategy an individual should adopt to gain the best possible outcome. In the dilemma, the participants are faced with the following outcomes:

  • If A and B each betray the other, each of them serves 2 years in prison
  • If A betrays B but B remains silent, A will be set free and B will serve 3 years in prison (and vice versa)
  • If A and B both remain silent, both of them will only serve 1 year in prison

Looking at this, it is evident that it would be best for A to betray B, while B remains silent. Option 3 would be the second best bet followed by 1 and the inversion of 2. How can we optimize our strategy in this dilemmic play?

In the 1970s, Robert Axelrod programmed the dilemma with different players and strategies and let them have a “tournament”. The outcome was that the “Tit for Tat”-strategy was the most successful and drove the other strategies into extinction. Applying that strategy, the individual starts out cooperatively. After that, s/he copies the other player’s decision (if s/he cheats, I will cheat; if s/he cooperates, I will cooperate).

This strategy is successful for the following reasons:

1. It is nice (starting point)
2. It retaliates if the other player does something negative
3. It is forgiving
4. It’s clear cut in its rules (not probabilistic)

The “Tit for Tat”-approach has a weakness, though. It is vulnerable to signal error. If individual A acts cooperatively, but the cheating signal is passed on, individual B will cheat in response which would lead into a spiral of competitive behavior.

To avoid this problem, the “Forgiving Tit for Tat”-strategy was introduced. It generally works the same way, but if the forgiving individual notices that the behavior has spiraled down to a succession of competitive behaviors, it switches back and decides to act in a cooperative way. This solves the first problem, but leads to a different one: The “Forgiving Tit for Tat”-strategy can easily be exploited.

As a result, scientists found that it is best to start off with the regular “Tit for Tat”-approach and switch to the forgiving version if, and only if, the other person has gone long enough without cheating to earn one’s trust.


How does this translate to animal behavior?

Pretty well, in some cases. Among others, Sapolsky references studies on female vampire bats that share their nests and feed all of the young according to the principle of reciprocal altruism. If one female cheats (or if it seems like it to the others), the other mothers will stop feeding this female’s young.

However, the more research was done and the closer the researchers looked, the more exceptions were found. It can plausibly be said that this is a result of the complexity that comes with real life situations. The individuals are not only engaged in one “game” with one other individual. They are interacting in a multitude of different situations with many individuals and often within a preexisting power structure. Those situations can not (only) be seen as singular events. Instead, one has to recognize that they are intertwined and the actions, reactions, and outcomes influence one another.