Amanda Garrett

Stacey Smith


Best, M. L. (1999).  How culture can guide evolution: An inquiry into gene/meme          enhancement and opposition.  Adaptive Behavior, 7(3/4), 289-306.



            In this research paper Michael Best seeks to explore the relationship between culture (social learning, also called “meme”), genes and individual learning in achieving adaptive goals.  The method is modeled after Hinton and Nowlan’s (1987) computer simulation looking at individual learning and evolution.  In order to study social learning, it is expressed in an imitation algorithm for the purposes of the computer simulation.   Best looked at situations where gene/culture was enhanced (i.e., both going toward the same adaptive goal) and when they were opposed.   He seeks to answer the following question: Can cultural influences guide evolution in the absence of individual learning?


Definition of Gene/Meme Enhancement

            In order to study genetic evolution and learning, we must first understand what Best means by “learning”.  He is using this term in reference to “lifetime learning” (which have links to genetic traits and evolution) versus “high-level learning” (i.e., philosophy).  More specifically, Best mentions memory, foraging, food preferences, mate selection, predator and risk avoidance as examples of learning.

            When both cultural and genetic selection favor the same trait, the combined force is much stronger than it would have been had only the genetic influence been present.  This is the definition of gene/meme enhancement.  Logically, the overall evolutionary adaptive goal is reached much quicker with both influences working together.


Past Research

            Baldwin argued that although learning was not passed down (heritable), learned traits could still have an impact on genetic evolution.  Hinton and Nowlan (1987) added weight to this claim by using a simulation program.  A genetic algorithm was used to express individual learning.  For example, each “agent” in the simulation is assigned twenty characters, which represented their genotype. These characters can take the values of “1”, “0” or “?”.  Through the program, each agent gets to participate in a “learning round” where the character that is filled with a question mark has a 50% chance of being replaced with a “1” and a 50% chance of getting a “0”.  Agents that end up with all “1’s” in their genetic makeup are the most genetically fit.  Then a genetic program was run to simulate each agent reproducing.  The reproduction was adjusted to favor the genetically fit agents (i.e., the agents all or mostly made up of “1’s”).  One would expect, if evolution was taking place in this simulated world, the percentage of “0's” to eventually “die out” and the percentage of “1’s” to increase.  This is exactly what happened.  Hinton and Nowlan (1987) found that when the program is run without learning there is no difference between the percentage of “1’s” and “0’s” in the population, thereby concluding that “in the absence of individual learning genetic evolution is lost.” 

            Belew (1990) argued that genetic evolution, individual learning and social learning all work as one adaptive system.  He said that successful agents transmit a cultural advantage or learning bias to their offspring.  He noticed that this transmission of an advantage allowed the population to achieve the adaptive goal faster.


The Interaction of Gene/Meme Enhancement and Social Learning

            Best replicated Hinton and Nowlan’s (1987) simulation, but he replaced individual learning with social learning.  He did this through the use of imitation, which he argues is the foundation of human culture.  Under his social learning model there is no random assignment of “0’s” or “1’s” to agents.  What takes place is a transmission of “0’s” and “1’s” from the model to the learner.  Another way of thinking about it is successful humans teach the adaptive behavior to many learners who will only know the “right” behavior, instead of having to individually learn through trial and error. What he discovered is that this social learning paradigm outperforms, or reaches the adaptive goal, quicker than individual learning.  This has occurred because one model is able to transmit the behavior to many learners, thus speeding up the process.

            He concluded that humans do not need individual learning in order to make evolutionary progress – social learning can fulfill this function on its own (and faster).  However, he also points out that no animal relies solely upon social learning.


The Problem of Gene/Meme Opposition

            The nature of gene/meme (culture) opposition is “when cultural selection favors memes which are not the most advantageous in terms of genetic inclusive fitness.”  To test the opposition, the author ran the same simulation, only he employed all three mechanisms (individual learning, social learning and genetic evolution).  He pitted genetic evolution and social learning against each other. In other words, they were striving toward two different adaptive goals (i.e., social learning toward and all “0” population and genetic evolution toward and all “1” population).  Individual learning is paired with genetic evolution to give it “a fighting chance.”  However, even with strong social learning forces, genetic evolution is hardly affected and always eventually wins out.  Best did acknowledge the fact that a strong opposing culture can “neutralize” the effects of genetic evolution so the simulation is in a draw, with neither mechanism able to reach their goal for the population.  By keeping some ground without any help, Best mentions that social learning, if aided in some way, might have the potential to win out in this situation.  It is needless to say that a weak opposing culture has no effect on the pull of genetic evolution.




            Best showed that when social learning is enhancing genetic evolution it can guide the process equally as well as, and more rapidly than, individual learning.  This implies that we do not necessarily need individual learning in order to make progress or reach an adaptive goal.   Since the imitation algorithm was so successful, it is speculated (with supporting past research) that imitation is the foundation of human culture.  However, when social learning is pitted against genetic evolution it is defeated.  The best a strong culture can hope for is a draw in the evolutionary game of tug-of-war.





I.                    Definition of Gene/Meme Enhancement

A.     Genetic and cultural selection both favor the same trait.

B.     The combined selective force is stronger than genetic influences alone.

C.     Adaptive goal is reached more quickly with both influences working together.

II.                 Findings of Past Research

A.     Hinton and Nowlan (1987) show that individual learning guides evolution.

1.      Used simulation program to show effects of individual learning   on evolution.

2.      Baldwin” effect – while not hereditary, learned traits still effect genetic evolution.

3.      Individual lifetime learning does interact with traits and evolution – such as the following traits.

a.       spatial orientation

b.      foraging

c.       mate selection

4.      Genetic evolution needs individual learning.

a.       Learning mechanisms increase chance of finding adaptive goal.

b.      Those close to adaptive goal will probably find it within their lifetime.

B.     Belew (1990) says that genetic evolution, individual learning, and social learning all work as one adaptive system.

1.      Successful parents vertically transmit learning bias to offspring.

2.      This allowed the population to achieve the adaptive goal in less time.

III.               The Interaction of Gene/Meme Enhancement and Social Learning

A.     Author uses same simulation as Hinton and Nowlan, but tests social learning and not individual learning.

1.      Imitation is the foundation of human culture.

2.      Successful humans teach the adaptive behavior to many learners who will only know the “right” behavior.

B.     Results indicate that social learning is superior to individual learning – it caused the adaptive goal to be reached faster.

C.     Humans do not need individual learning in order to make progress towards adaptation as long as they have social learning.

D.     No animals rely solely upon social learning and do not use individual learning.

IV.              The Problem of Culture and Gene Opposition

A.     When culture selects genes that are not advantageous for inclusive fitness.

B.     Genetic evolution will eventually win out over culture/social learning.

1.      Strong opposing culture can neutralize effects of genetic evolution.

2.      Weak opposing culture has no effect on evolution.

C.     Culture/social learning, when enhancing genetic evolution, can guide evolution.

V.                 Conclusions

A.     We do not need individual learning in order to make progress towards an adaptive behavior.

B.     Imitation is the foundation of human culture.

C.     Genetic evolution will ultimately win over culture.