• this book is really well written. The ideas aren't that complicated (basically competition is between genes + game theory/stable strategies are the two main ideas), but the explanatory power is huge

  • Many people conceptualize the competition of evolution as operating at the level of individuals or even species, but Richard Dawkins argues that evolution at the level of the gene makes a lot more sense

    • the first chapter or two argues why this makes sense, and the rest of the book is reasoning about what this would predict in different cases, contrasting with other explanations and providing some evidence

  • New story:

    • genes (some segment of a chromosome) are competing for space in the gene pool against their alleles.
    • Genes create survival machines (bodies)
    • genes influence the behavior of animals by baking in patterns. Some of which include learning or are otherwise sophisticated

  • Perspective on aging

    • aging isn't a "pre-programmed decline," but selective pressures mostly care about pre-reproductive success and a little bit of post-reproductive success. So if there are genes that help a bit in that period, but happen to hurt after that period, they'll still be selected
      • I'm not sure if I'm totally convinced by this

Evolutionary Stable Strategy

An Evolutionary Stable Strategy is essentially a {1:Nash Equilibrium}. There are some strategies that, if developed by a majority of a population, {1:penalize individuals from deviating}

  • Of course, this normally doesn't achieve the species-wide optimum

At the level of individuals, it can be used to understand inter-individual behavior, ex. that many animals take a "gloved fist" approach to aggression.

When there are asymmetries, the story is a little more interesting.

  • Lions chase antelopes and antelopes flee instead of fighting back due to the fighting ability asymmetry.
    • this concept can also be used to understand how genes are selected against the backdrop of other genes in a body. If a gene pool is dominated by herbivorous genes, then a gene for sharp teeth will probably be penalized, even though it's not a "bad gene" in general.

Kin Selection and Altruism

The Selfish Gene

when should we expect to see altruism (the sacrifice of some resources for the self for others)? #card

  • when the benefit to the other times their genetic relation outweighs the benefit to oneself times 100%.

  • genetic relations:

    • parent-child, siblings: 50%
    • aunts & uncles, grandparent-children: 25%
    • etc.

  • this is complicated by the fact that it might be hard to know for certain how related you are to someone

    • there are cases where the EV for relatedness to nearby animals of the same species is high enough that it makes sense to be generally altruistic (ex. calls to alert predators)

  • Parental care is a special case of kin selection. Theoretically an older sister shouldn't distinguish between a baby brother and a son, except for the fact that she may be more certain the son is hers

  • Why don't animal populations keep increasing? #card

    • there is an optimal litter size that maximizes the number of grandchildren that will live
      • it might be variable - ex. female mice have smaller litters when in an over crowded area
    • menopause - one explanation is that it becomes better for the grandmother's genes to stop bearing new children and invest in grandchildren instead

  • there is a mismatch in how much the parent vs. child want child care

    • the parent invests resources in children based on their life expectancy and ability to care for themselves
    • there comes a point when it's in the interest of the parent's genes to switch from {1:caring for that child} to {1:caring/bearing younger children}, but the child, which is {1:100%} invested in itself and {1:50%} invested in its siblings, has a lag before it wants the same thing
      • this leads to the child lying and generally trying to get more than its "fair share", at least for a time

Battle of the Sexes

  • You can also reason about many questions about mates using the selfish gene framework. Both parents want the other to invest more in the child

    • What distinguishes males from females?
      • mainly, that one gamete (sex cell) is bigger than the other. Essentially an Evolutionary Stable Strategy is one part of the species getting locking in to larger and larger gametes (which becomes the females/egg-producers), while the other half exploits this more and more, producing many small mobile gene delivery gametes that don't provide nutrients
    • ^ can be seen as the beginning of males' exploitation of females. While each parent has a pressure to get the other parent to invest more resources in the child, the female has already invested more, so ex. if she abandons the child to the father, it's easier for the father to retaliate by abandoning it himself (and find another female that won't abandon).
    • female strategies
      • he-man strategy
        • don't care about the father investing, but select for very fit fathers. Anything that gets selected for (ex. long, colorful tails) will soon to be selected for strictly because other females select for it
      • domestic-bliss strategy
        • attempt to select a father who will stay and invest resources. This can be forcing the father to wait a long courtship period or build a nest or something
    • why a sex ratio of 50/50?
      • Just a few males can seed the next generation, so 50% might seem excessive. But when you view this as a selfish gene decision from the parent perspective, as soon as the ratio tips, ex. towards more females, there is a pressure to have more males as they will be able to seed more of the next generation (and when it tips towards males it makes sense to have more females)

  • Other interesting behaviors

    • always at least one selfish gene explanation for altruism
    • herding
      • Simple model (there are more sophisticated ones too)
        • if predators go for the nearest prey, each prey has a domain of danger around them they want to minimize, so they selfishly aim to be in the middle of a group, which leads to the emergent behavior of bunching / herds
    • bird calls to alert that predators are nearby
      • these bird calls are designed to be hard to locate (really cool) - so that means they're inherently risky for the giver
      • many potential selfish-gene explanations, but here's one
        • when a bird sees a predator, they want to fly away, but it's risky to be the only one, so the call helps them leave as a group
    • gazelle stotting - also called pronking :)
      • seems altruistic, but one explanation is actually that the gazelles are signalling to the predator, "look how high I can jump, I'm a healthy and fast gazelle, so you should chase someone else"
    • Social insects
      • cases where certain organisms are sterile, which leads to interesting behavior (ex. suicidal "altruism")
      • in all cases except termites, this only happens with Hymenoptera, in which males are given one set of genes
        • this changes the relatedness calculations, and can make sisters more related to each other than the mother is to them (on average)
        • if you do the math, the optimal sex ratio for the mother is 1:1, while it's 3:1 for the workers (daughters only? or males too?)
          • we expect the workers to win because they have more control, and indeed a 3:1 ratio has been observed for several species
            • interestingly, for slave-taking species (where workers are actually unknowingly slaves taken from others species as eggs), we would expect the ratio to be 1:1, and it does seem to be
    • symbiosis
      • cleaner-fish have a territory where bigger fish come for repeated cleanings - this probably makes it worth it to not eat them after they clean you
      • some ants and termites farm fungi! And use aphids to better suck juice from grass - symbiotic because the ants take care of everything
      • classic mitochondria inside us was/is symbiosis?
      • lichen is actually a fungus + an algae (and maybe yeast??)

Nice Guys Finish First

in a single round of {1:Prisoner's Dilemma}, it doesn't make sense to cooperate, but there are many circumstances in nature that are really {1:iterated Prisoner's Dilemma}, which has very {1:"nice"} winning/stable strategies

  • the example he uses is birds picking some bad bug off each other's heads (where they can't reach themselves), with the assumption that they can recognize individuals and thus know who has helped vs. taken advantage of them
  • against a wide array of opposing strategies, certain traits emerge as "winners", in terms of maximizing total reward
    • nice - starting the first round with cooperate
    • retaliatory - responding to some amount of defections with one or more defections
    • forgiving - cooperating once again it the other player doesn't continue to defect
  • a classic example is "Tit for Tat", in which you simply start with "cooperate" and then play the last thing the other player played
  • both something like "Tit for Tat" and a "always defect" strategy are "stable" in a population, in that when a population reaches a critical mass of them deviation is punished
    • but if you have a local neighborhood (ex. if organisms you're more related to are close by) of "Tit for Tat" in a larger population of "always defect", they will do better than the larger population. This isn't true for "always defect", so it seems like nicer strategies have a higher level stability