Simulation of evolution

Simulation of evolution | www.agner.org

 
thread incomplete modeling - Carmi - 2003-04-30
last replythread incomplete modeling - A Fog - 2003-04-30
last replythread incomplete modeling - Carmi Turchick - 2003-05-03
last replythread incomplete modeling - A Fog - 2003-05-03
last replythread incomplete modeling - carmi turchick - 2003-05-06
last replythread incomplete modeling - A Fog - 2003-05-06
last reply incomplete modeling - Carmi Turchick - 2003-05-07
 
incomplete modeling
Author:  Date: 2003-04-30 04:56
The issue of altruism vs. egoism, and how to accurately model it to reflect the real world, is one which I find frustrating due to what I percieve as some very basic flaws in the models I have seen to date, including the one I see here. Humans have a very well developed faculty to detect cheaters. Why? Humans have a very strong predisposition to scorn whistle-blowers and tattle-tales and those who are not "team players". Humans are very likely to exert peer pressure on those not conforming to the norm, to "pick on" those who are noticeably different. In short, humans are very strongly disposed to making sure there is conformity and group cohesion and allegiance and no "traitors". In real world terms this means that, for example, a strict egoist moving into an altruistic tribe would not fare well at all. In fact the egoist would be very likely to wind up an outcast or dead within a very short period of time. Neither of which would be an evolutionary advantage.
Since tribes of altruists will prove stronger than tribes that allow even a few egoists, altruism wins. This reflects the real world as well, at least it explains the prevelance of altruistic behavior towards non-kin.
   
incomplete modeling
Author: A Fog Date: 2003-04-30 09:14
I have indeed included a model of coevolution of altruism and conformity. The studies of this model is not finished.

In the literature on evolutionary psychology you will find a lot of theorizing about detection of cheaters, predicting when cheating is likely to be detected and when not, etc. All inspired by game theory.

   
incomplete modeling
Author:  Date: 2003-05-03 01:53
From what I have seen of Game Theory various methods are used to assign point values to certain strategies based on the Prisoners Dilema base method, mutual cooperation being good, mutual defection being bad and 50/50 being good for the defector while horrible for the cooperator. It is all well and good but it does not hold up even in long term strategy games, one of which I played for ten years. Long term strategy games showed me that defecting is a quick path to doom, as many players gang up on you and wipe you out. In a EEA scenario (Environment of Evolutionary Adaptation) defecting is a dead end geneticaly. If an individual defects and is detected they are likely to at the very least fall to the bottom of the status ladder, as present day criminals do generaly. There is also a very real chance of being slain in some painful way or being banished which pretty much always lead to death as well. Defecting without being detected can be advantageous, but the tribe that is not adept at finding defectors will find that it is weaker than its conspecifics who do find defectors well, resulting in the whole tribe being wiped out or at least the men killed with the women taken by the victorious altruists.
In short my problem with Game Theory is that I fail to see it accounting for the fact that life is very different from a computer game where losing is easily remedied. Even a few defections can easily kill in life and there is no replay button. Defecting is a very high risk strategy and not one that will be selected for generaly. A group of defectors has no chance against a group of altruists. Defectors cant be entirely eliminated, but they are sharply curtailed by any thriving group.
   
incomplete modeling
Author: A Fog Date: 2003-05-03 13:49
Carmi Turchick wrote:
Defecting without being detected can be advantageous, but the tribe that is not adept at finding defectors will find that it is weaker than its conspecifics who do find defectors well, resulting in the whole tribe being wiped out or at least the men killed with the women taken by the victorious altruists.
That's the problem: The captured women can carry the egoism gene into the victorious tribe.
   
incomplete modeling
Author:  Date: 2003-05-06 02:30
Yes, that could happen in a mathematical model. But try to think about the actual conditions our ancestors lived in instead. Members of a tribe were highly interdependant on each other for survival. Specialization of skills and the fair exchange of, for example, well made axes for applied knowledge of medicinal plants, was crucial for the survival of the individuals and the group. Sharing of food, a very basic foundation of human bonding in all known societies, was also crucial for survival. A tribe of egoists would see members who fell ill even from common illnesses which one can recover from in a week die instead of recover. They would not survive the week without food. Intra-tribal conflict would be nearly constant with a high resulting mortality rate. Those who failed to make a kill or went foraging in a desolate area would go hungry. There is a huge list of activities that our ancestors carried out that a tribe of egoists could not or would not, but which allowed us to survive. Survival of a tribe was far from a given and many perished without any intervention from an attack by another tribe. More than a few egoists in a tribe could easily select it out of the evolutionary game. A tribe of all egoists would not survive long enough to be wiped out in warfare and have its women taken as sex slaves. Children would not thrive or survive to adulthood as only their own parents would protect them, and if the father was hunting and the mother fell ill then the child would be doomed.
There are lots of ways in which the very concept of a tribe of egoists can be shown to be absolutely absurd given that this tribe would actualy have to eat and heal the sick and raise the young and make shelter and protect itself from predators.
We are interdependant and therefore egoists are not the norm and we have evolved many systems to detect them and stop them.
   
incomplete modeling
Author: A Fog Date: 2003-05-06 07:04
carmi turchick wrote:
We are interdependant and therefore egoists are not the norm and we have evolved many systems to detect them and stop them.
And we have evolved many ways to avoid detection of cheeting and to change strategy depending on the situation. In your rosy scenario there would be no thieves, robbers, murderers, or cunning businessmen selling trashy products at elevated prices.
The theoretical study of the neverending competition between altruism and egoism is a fascinating one with endless possibilities.
   
incomplete modeling
Author:  Date: 2003-05-07 04:33
C'mon, I made it abundantly clear that I was not suggesting the existence of tribes that are composed of people who are 100% altruists 100% of the time.
Of course I recognize the dialectical evolution between those attempting to enforce altruistic behavior and those attempting to evade detection.
Thanks for reminding me of another point I had failed to make though.
Most models dont account for what most acts of egoism are, infrequent acts of opportunity, and for the duality that exists in us all to behave both altruisticaly and egoisticaly. I wish the book tittle was coming to mind but there is a fascinating book where the author interviewed 50 people on Death Row about their views of themselves. They uniformly see themselves as basicaly good, as altruists, despite their murders and rapes and etc. One inmate recounts how he went shopping for the old woman upstairs who could no longer make it to the store and on his way home raped and killed a teenage girl.