Agner Fog: Cultural selection © 1999

3. Fundamental model for cultural selection

3.1 The genetic basis of culture

Cultural evolution is much faster than genetic evolution, for reasons that will be explained later. This has given humans an enormous advantage over other animals in terms of adaptability. The human capacity for culture is based on our genes. You may call it a metaadaptation because this trait is an ability to adapt rather than an adaptation in itself. In this way, you may say that genetic evolution has created its own replacement. (This is called vicarious selection, see chapt. 3.8)

Biological learning theory talks about programmed learning, by which we mean a pre-programmed faculty of learning a certain ability. The genes determine the ability to learn and set the limits for what may be learned and what may not. These limits may be wide or narrow (Gould, J. & Marler 1987; Mayr 1974; see also chapt. 12). The human capacity for culture is a learning program with very wide limits. The culture is not directly determined by the genes, and hence cannot be studied with the same methods as instinctive behavior.

An important aspect of the capacity for culture is the ability to learn from conspecifics. Learned behaviors may be acquired by observation and by listening to spoken instructions. But the spoken language is not the only channel of information. Humans have a peculiar propensity for religion, supernatural beliefs, rituals, music, dance, etc. These seemingly irrational phenomena are important media for the transmission of arbitrary behavior patterns, rules and norms. They are so to say the genes of culture, as we will see later.

Another important aspect of the capacity for culture is our group behavior. Humans have a remarkable ability to collaborate and propensity to identify with a group; to classify other humans according to their group affiliation; to favor and overestimate members of one's own group and depreciate and discriminate against members of foreign groups (ethnocentrism); and to display ones group affiliation by means of body adornment, language, rituals, etc. (Hogg & Abrams 1988). This pronounced group behavior means that the social group plays an important role in the cultural selection process, as we will see in chapter 4.


3.2 Cultural selection

Cultural selection theory is a theory about phenomena which can spread in a society, such as a religious ritual, a genre of art, or a certain fishing method. The theory entails three basic processes. First the phenomenon has to arise. This is called innovation. Next, the phenomenon may spread from one human to another or from one group of humans to another. This is called reproduction or transmission or imitation or diffusion. The third fundamental process in the theory is selection. By selection we mean any mechanism or factor that have an influence on how much or how little the phenomenon will spread. The most obvious kind of selection is the conscious choice exerted by humans.

Take agriculture as an example. For agriculture to start, some bright person must have invented a method of growing grains or other crops (innovation). This practice can then spread if the inventor gives on his idea to others who imitate his method (reproduction). Several conditions must be satisfied for this practice to spread effectively. Firstly, agriculturalists must be willing to give forth their knowledge to others. Secondly, non-agriculturalists must have contact with agriculturalists. Thirdly, non-agriculturalists must be willing to change their way of living. And finally, agriculturalists must be able to feed and raise a sufficient number of children. These four factors constitute the selection process, which is crucial for agriculture to spread in a population.

In this example, it is likely that the innovation has been made countless times without spreading, simply because it has been too labor-intensive. It took less time simply to collect the fruits that nature produced of itself, rather than growing them. In other words, the selection factors were not favorable for agriculture to spread until the natural resources became insufficient due to overpopulation (Rosenberg, M. 1990).

This model of cultural selection very much resembles Charles Darwin's theory of natural selection. The three fundamental processes are the same: variation (innovation), reproduction, and selection. The difference is that Darwin's theory is about genetic inheritance, while the cultural selection theory deals with cultural inheritance. Despite the formal similarity between the two theories, there are important differences which mean that you cannot draw any conclusions from one kind of selection to the other (Daly 1982). An important difference is that the cultural reproduction is not necessarily connected with human reproduction. A habit can not only be transmitted from parents to their children, but also to other humans unrelated to the inventor. The custom of living in convents and monasteries, for example, has spread despite the fact that monks and nuns do not have children. Another important difference between the two processes is that acquired traits can be transmitted by cultural inheritance, but not by genetic inheritance.

It may be instructive to imagine a cultural phenomenon as an independent living being, which like a parasite or virus can spread from human to human and reproduce independently of human reproduction. In fact, the cultural selection theory is a very good model for describing the spread of infectious diseases. It is important to realize that a cultural phenomenon may have its own 'interests' which are not always identical to the interests of its human bearer. In the agriculture example, this custom only spreads when it serves the interests of the humans. Agriculture's interests are identical to the humans' interests - or it would not spread. But in the disease example there is a conflict between the interests of the virus and its bearer. Another interesting example is drug addiction. Heroine can make a human do things that he or she normally would not do, including recruiting others to the drug in order to finance one's own consumption. The drug has its own 'will' which to some extent can override the addict's will. Heroine addiction can spread despite the fact that probably nobody wants to be a junkie.

We can discuss ad nauseam whether cultural selection should be seen as selfish 'culture-genes' which like parasites cynically manipulate humans in order to reproduce themselves, or it is intelligent humans with a free will who consciously choose those cultural forms that best serve our needs. This debate seems fruitless to me since the end result always will be the same: Certain cultural phenomena are copied and transmitted more than others, and the characteristics of each cultural phenomenon are just as decisive for the outcome of this selection process as are the characteristics of the humans. The two extreme standpoints mentioned above are nothing but opposite ways of viewing the same events. When you pin it down to mathematics, the end result will be the same no matter which side you are seeing the process from.

In the following paragraphs, I will define the unit of selection and the three fundamental elements in cultural selection theory: innovation, reproduction, and selection.


3.3 Unit of selection

If you want to set up a detailed model for a cultural selection process then you must define what is selected. It would be incorrect to regard human individuals or groups as units of selection since a person can change his/her cultural traits several times in the course of his lifetime. Generally, you have to define a cultural phenomenon, an idea, a conception, a way of thinking, a behavior pattern, an artifact, or the information that creates or controls such a phenomenon, as the unit of selection for the model in question. Certain cultural phenomena, such as clothing fashion, are carried by individuals, whereas other phenomena, such as the form of political organization, can only be carried by a group of people in union.

In genetic selection models, the genes are often treated as indivisible units, but cultural units of information are seldom indivisible. If, for example, a certain piece of music tops the hit lists, then it is the piece of music as a whole which is selected, but if another composer gets inspired by this music and incorporates a certain sound or chord from it into a new piece of music, then it is only a little part of the original piece of music that is selected. The selection takes place at many levels, and you should not be misled by the term 'unit of selection' to think that it is something indivisible.

The difficulty in defining a cultural unit of selection has often led to confusion. You have to find a chunk of information which sticks together long enough for its spreading and selection to be measured. How big this chunk should be depends on the phenomenon you want to analyze. In principle, a complete analysis would have to comprise all levels of selection (e.g. the musical genre, the melody, a motif, a particular chord, and a particular sound), but for practical purposes it will often suffice to concentrate on one particular level which fits the purpose of the analysis. When studying a higher level in the hierarchy of information, such as an entire religion, a school of thought, or an art genre, then it is most appropriate to call it a meme complex rather than a single meme.

The problem with selection at multiple levels also exists in genetic evolution, where genes are seldom perfectly indivisible, and where not only genes and individuals may be selected, but also families of related individuals, groups of individuals, and even entire species and clades. But the problem is possibly bigger in cultural selection theory because the units are more amorphous and the levels not sharply distinguishable. Even in a fuzzy system where all cultural units can be divided into subunits, valid units of selection can still be found by statistical techniques (Pocklington & Best 1997).

There have been many attempts to define a universal unit of selection, and these definitions have led to endless controversies in biological as well as in cultural evolution theory. One of the most general definitions is a replicator: Anything that is replicated relatively unchanged can count as a unit of selection (Hull 1988). This definition is not perfect, however. Consider the situation where innovations are judged by the inventors as good or bad. Good innovations are shared with others and replicated, whereas bad innovations are rejected by the inventor without telling anybody. This is obviously a selection process, and indeed a very important and powerful one. But if you define a unit of selection as a replicator then you have a problem because the de-selected innovations are never replicated and hence do not count as units of selection. You become unable to define the selection process unless you include potential replicators in the definition. But since almost everything in the world are potential replicators, this definition becomes over-inclusive and therefore useless. In my opinion, there is no universally useful definition of a unit of selection. You will have to choose from time to time the definition that best suits the phenomenon you want to study. In most specific cases of cultural selection you will find that the choice of unit of selection is quite obvious and uncomplicated, and that you can easily make a useful analysis of a specific phenomenon without having a definition that also applies to all other phenomena.

Biological theory makes a distinction between genotype and phenotype. Certain scientists make a similar distinction in the theory of cultural reproduction. Sociobiologists have defined a cultural disposition or a cultural unit of information, and several names have been proposed for such a unit of cultural information: artifact type, mnemotype, idea, idene, sociogene, instruction, culture type, culturgen, meme, menteme, concept, rule, mental representation (Lumsden & Wilson 1981; Stuart-Fox 1986; Hill 1989, Burns & Dietz 1992). The word meme is the name that has caught on, and the analysis of memes is called memetics.

The external manifestation of a meme can adequately be designated by the same word as in genetic theory: phenotype, in particular when connected to individuals. The reason for using the same word in both cases is that the appearance or behavior of an individual is usually determined by both genes and culture, and it is usually difficult to separate these two factors. In more general terms you may distinguish between replicators and interactors. An interactor is the functional expression of a replicator that causes the selection (Hull 1982, Speel 1997).

Few sociologists have made a similar distinction between cultural dispositions and their external manifestation. In his sociology of education, Pierre Bourdieu makes a distinction between the inner cultural dispositions of people, which he calls habitus, and the external manifestation, called social structure (Bourdieu & Passeron 1970, Bourdieu 1979).

This distinction between a cultural unit of information and its manifestation is sometimes necessary. For example, it would be incorrect to say that a stone axe can reproduce itself. What is reproduced is the recipe for how to make a stone axe. The recipe is the reproducible unit of information, and the axe its external manifestation. The human who makes the stone axe is the bearer or host of the meme and the vehicle for its transmission. In other cases the distinction is unimportant, for example in the case of story telling (Ball 1984).

The unit of selection may be regarded as qualitative (e.g. whether people go to church or not), or quantitative (how often people go to church). Quantitative traits cannot be described adequately in terms of atomic units such as memes - we need a different kind of model. The following example will illustrate this:

In a free market economy, big companies tend to grow even bigger. This is because big companies enjoy the benefits of rationalization, mass production, mass distribution, mass advertising, synergism effects, and less competition (see chapt. 5.12). The mechanisms that favor big companies over small ones obviously constitute a selection process, but not one that can be described completely in terms of memes. You have to define the sizes of the companies in quantitative parameters and analyze the dynamics of feedback in the system. Such models are well known in mathematics and adequate methods exist for their analyses (e.g. differential equations or laplace transforms).


3.4 Innovation

The cultural analogue to mutation is called innovation. A cultural innovation can be a new idea, a new way of obtaining food, a new ritual, a new song, a new rule, a change in the social structure, etc. When the word variation is used in this context it is a relic from Darwin's time when it was not known how new variants arose. Since variation usually designates a state of differentness in a population, it is reasonable to regard innovation (and mutation) as a more fundamental concept than variation, and the two concepts should not be regarded as synonymous.

The word innovation should not mislead anybody to think that this necessarily means rational and ingenious inventions. The word is used regardless of whether the new form deviates much or little from the previously known forms and whether the new phenomenon has arisen accidentally or as a result of intelligent thinking.

In biological evolutionary theory, all mutations are generally believed to be blind and random. But in the cultural process innovations are seldom completely random. Innovations may very well be planned, and they are often advantageous to the inventor.

Some inventions occur accidentally and randomly, for example by play or blind experimentation (Vandenberg 1981). If such an experiment turns out to have (subjectively) advantageous consequences, then it will be repeated, and a discovery has been made. But not all innovations are spontaneous and random discoveries. Most innovations are provoked by a certain problem that people attempt to solve. They may be ingeniously thought out or random experiments, but very often with the conscious aim of solving a certain perceived problem.

Innovations are not necessarily advantageous to the inventor. An innovation which has been made with the best intentions may have unintended consequences which makes it less advantageous than expected. Innovations can also occur by imitation errors. If a person has forgotten the exact procedure for doing something, and hence does it imperfectly, he may pass on an inferior version to his successors. This is a disadvantageous innovation.

Religion is an area where blind or irrational inventions readily occur. Even a seriously mentally ill person can be a prophet if his social group accept his hallucinations as revelations. But mental illness is not a satisfactory explanation for deeply irrational decisions. Many societies have religious rituals which purposely induce hallucinations. This can be done in several ways: by hallucinogenic drugs, meditation, self-suggestion, ecstasy, sensory deprivation, etc. These hallucinations are interpreted by the religious actors as visions, revelations, prophecies, or omens, and they act in accordance with this interpretation. In lack of hallucinations, other means like interpretation of dreams, the weather, or other random occurrences, may of course be used in a similar way. But neither dreams nor hallucinations are truly random. They are products of the unconscious mind, which in turn is influenced by a lot of cultural phenomena such as religion, art, rituals, myths, emotional interactions, etc. The interpretation of these hallucinations, dreams, etc., is also to a high degree a matter of choice. The actors often consciously or unconsciously choose among the possible interpretations the one that gives the most advantageous conclusion. In this way a seemingly irrational hallucination may be a catalyst for a more or less intelligent decision.

Even in the most secular societies, some people make prophecies and omens based on random occurrences such as the position of stars and planets, the movements of birds and animals, the lines in the hand, coffee grounds, playing cards, or dice. It must be said that in the lack of positive knowledge people have a marked tendency for generating fictive knowledge and to act accordingly.

An innovation may be done in vain if the selection conditions are unfavorable at the time. Such an innovation may be forgotten or it may be saved in society's 'knowledge bank' to be later activated should the selection conditions change. A pluralistic society may in this way store an immense reserve of cultural possibilities, lying latently as unrealized ideas or as outmoded customs and rituals which are only upheld by a few deviants, but which may later come out and be activated or reactivated if altered selection conditions should favor them.

An innovation may even lag behind the development which it seemingly gives rise to. I am here particularly referring to moral and religious systems which may be ex post facto rationalizations which justify or stabilize a structure that already exists or is under construction.

We may conclude that there are rich possibilities for both rational and irrational, advantageous and disadvantageous, innovations to occur in any society. In a pluralistic society you may find proponents of practically any point of view. Such a society will have an almost inexhaustible reservoir of ideas and cultural variations to choose between. Selection decides which of these innumerable ideas are promoted at which time.

Social innovations are often regarded by society as deviations, and the persons representing them are sometimes persecuted as deviants. The oppression of deviants has the function of preserving the social system. If a deviation spreads and wins acceptance in significant parts of the society despite all attempts at suppression, then we may say that a social change has occurred. A well-documented example of a religious deviation which has had a significant influence on the development of a society, has been described by Lewellen (1979). The significance of deviations will be further explained in chapter 8.


3.5 Reproduction

Reproduction of cultural traits is the second of the three elements in cultural selection theory, and the element that has been most thoroughly studied. Transmission of cultural traits can follow different patterns: vertical transmission is from parent to child; horizontal transmission is between unrelated persons; group socialization is the concerted influence of many (older) group members on a child or new group member; and the one-to-many transmission pattern is the influence of a teacher or leader on a group (Guglielmino, 1995).

Theories about how a human acquires knowledge, skills, norms, beliefs, and attitudes from others are known from the psychology of learning, biological theory of learning and imprinting, sociology of education, socialization theory, etc. This area has been so thoroughly covered by several disciplines that it is superfluous to go into details here.

I will only mention a single kind of cultural reproduction which has been insufficiently studied, namely unconscious communication. All societies are filled with irrational phenomena such as art, music, dance, religious rituals, myths, stories, and so on. The possibility of an unconscious communication or transmission from generation to generation through such media is included in the psychoanalytic theory of Freud:

"For psycho-analysis has shown us that everyone possesses in his unconscious mental activity an apparatus which enables him to interpret other people's reactions, that is, to undo the distortions which other people have imposed on the expression of their feelings. An unconscious understanding such as this of all customs, ceremonies and dogmas left behind by the original relation to the [primal] father may have made it possible for later generations to take over their heritage of emotion." (Freud 1913:159)8.

If we refuse to accept that a rain dance really can make rain, then we must assume that the dance has some other function such as creating group solidarity or communicating some other values and emotions. The participants may not consciously realize this function, so there is reason to assume that the unconscious plays an important role here. My postulate is that many irrational human activities are media for a communication from the unconscious of one person to the unconscious of another. The persons involved may not realize that this communication takes place. I will support this theory later, especially in chapter 11. These irrational phenomena may be regarded as functional in the sense that they influence human behavior in such a way that they promote their own reproduction.

A tale about good and wicked people, where the wicked people die and the good people become happy in the end after many troubles and torments, may have the function that it conveys the norms of good and bad, and tells people not to give up when times are hard. Perhaps neither the storyteller nor the listeners are aware that this norm-communication takes place. Their conscious motive for telling or listening is a very different one: that the story is exciting. The criterion for what is experienced as exciting is probably a result of genetic as well as cultural selection.

A ritual dance performed by a shaman may tell his tribe: I have magic powers and I am your spiritual leader. A chain dance where everybody participates communicates values of solidarity and unity, while ballroom dancing where men and women dance in pairs sends the message that society is built upon the nuclear family as the basic unit.

The consequence of such a widespread unconscious communication is that every society has a collective mass of unconscious messages that influence and homogenize their behavior. Such unconscious communication is very difficult to study because of the tremendous methodological problems in the science of psychoanalysis. There is nevertheless general agreement that the unconscious has an important influence on behavior. How such unconscious communication takes place is therefore certainly worth studying. I will return to this question in chapter 11.


3.6 Selection

The most fundamental principles for cultural selection were implied in the agriculture example on page 60. Three basic requirements must be satisfied for a cultural phenomenon to be transmitted from one person to another: Firstly, the two persons must have contact with each other; secondly, the sender must be willing to share his knowledge or unable to hide it; and thirdly, the receiver must be willing to accept this knowledge and to incorporate it into his behavioral repertoire. If one or more of these conditions are not fulfilled, then the phenomenon will be eliminated. The requirement for contact between sender and receiver means that such customs which imply social isolation do not easily spread. Selection by the emitter of information causes the spreading of such religions which command their believers to proselytize. Selection at the receiving end is seen in the trivial fact that if the advantage of a technological innovation is obvious to everybody, then it will spread fast.

A more primitive selection mechanism is the selective survival and reproduction of individuals based on their culturally determined behavior. An extreme example is certain religious sects which totally ban sexual intercourse, such as the Shakers. These sects have of course become extinct for lack of descendants. But celibacy memes are not always unfit. An individual who does not have to care for spouse and children will have more time and energy for preaching and spreading his memes (Ball 1984).

Religious prohibitions against contraception and abortion, on the other hand, is a mechanism which increases the number of genetic offspring - and hence also cultural offspring.

Selective survival may also operate at the group level. When one people defeats another in war, then the loosing group is seldom completely exterminated, but their religion and their political organization may very well be eradicated when the victors force their religious and political principles on the losers. Whether a tribe is destroyed or its religion is annihilated is admittedly two very different occurrences, but the effect in terms of cultural selection is the same: that those religions and those forms of political organization which generate military superiority will advance at the expense of weaker strategies.

An interesting kind of selection takes place when the receiver chooses between alternative senders. This is selective imitation. It is well known that successful businesspeople, athletes, or artists are imitated more often that those who are fiascoes. Other examples of selective imitation are teenagers' hero worshipping, or the imitation in developing countries of everything that belongs to the western industrialized world.

Boyd and Richerson have set up a model for this kind of selection under the name of indirect bias. The characteristic that makes certain people the preferred models for imitation are called indicator trait. And the criteria for choice of imitation model in the persons who imitate, is called preference trait. If both indicator trait and preference trait are subject to cultural selection and these two selection processes are coupled, then the process may run away and lead to exaggerated displays of prestige, according to Boyd and Richerson (1985).

Another interesting selection mechanism is the selective promotion of certain persons to influential positions such as leaders or educators (Campbell 1965).

More well-known mechanisms are economic competition and democratic elections.

Psychology often plays an important role in the selection of memes. The fittest memes often have a strong psychological appeal. They push our buttons. A psychological button is a metaphor for the stimulus/response mechanisms that make us pay special attention to certain topics, such as danger, food, and sex. These are topics that have been of vital importance throughout our evolutionary history, and a strong selection pressure has always existed for paying attention to any information about these subjects. Some of the most effective buttons are described on page 44.

Take jokes as an example. A joke about danger or sex pushes our buttons. This makes us pay attention to the joke, remember it, and pass it on to others. Jokes that do not push any buttons are soon forgotten, whereas the button-pushing jokes are remembered and passed on. They have higher fitness.

Button pushing plays an important role in the competition between commercial TV channels because it attracts viewers. News stories about horror or sex pushes more buttons than other, possibly more relevant, topics. Whether the stories are true and accurate has little importance: truth is irrelevant to the fitness of a story as long as no proof or disproof is presented.

Button pushing is widely used in commercial advertising and political campaigning to make us pay attention. A product that is supported by button-pushing commercials can easily outcompete one that isn't (Brodie 1996).

A meme has to make sense and be easy to understand. A meme that is difficult to understand or incompatible with existing memes does not easily catch on. But something that does not make sense is good at attracting attention to a meme. A paradox or something that contradicts previous knowledge creates cognitive dissonance. Our minds will always struggle to make things make sense and thereby pay attention to whatever creates the cognitive dissonance as well as to the meme that seems to solve it. People are more receptive towards new memes when they are placed in a paradoxical or mentally uncomfortable situation such as a painful initiation ceremony (Brodie 1996).

A consequence of this is that a meme may catch on simply because it is remarkable. If, for example, a minority of the population spells a particular word in an awkward and complicated way while the majority spells it in a straightforward way, then readers are likely to notice the awkward spelling when they see it, but fail to take notice when they see the simple form. Hence, the awkward form may gain wider usage simply because people remember only the form they took notice of. This is called the bandwagon effect.

The fastest and most obvious form of selection is conscious choice and rational planning. In this context it is necessary to distinguish between the cultural information (meme) and the utilization of this meme in a specific action. Note that in the genetic selection theory the genes are tied to distinct loci, and one allele cannot enter a locus without displacing another allele, which is lost in the process. This is not so in cultural selection. One meme can enter the mind of a person without displacing any other meme. The person simply acquires new information without forgetting the old one. A human can thus have the knowledge of several different behavior patterns at the time he makes a choice. All the alternative memes are present in his brain, but only one of them is activated.

Many scientists have mistakenly described the choice of humans as a selective transmission of information, where it would be more correct to talk about selective use of the transmitted information. When, for example, the ladies' fashion recurrently changes between long and short dresses, then it would be absurd to claim that the information about short dresses displaces the information about long dresses, or vice versa. The women have always known everything about both long and short dresses. What is selected is not knowledge, but preferences. The fact that school children today learn about ancient religions without practicing those religions, can be taken as a proof that cultural information can be transmitted and preserved indefinitely, and still remain inactive. The too often ignored distinctions between knowing a meme, discussing it, endorsing it, and translating it into action, has been called levels of retention (Speel 1997). There are widely different ways of believing something and different degrees of commitment to a belief (Sperber 1990).

As already explained, both the innovation process and the reproduction process may be selective. In some cases it may therefore be difficult to separate innovation, reproduction, and selection as three distinct processes. Some theorists hold that cultural information is transformed or modified each time it is copied, and the accumulated effect of multiple such transformations makes the process tend towards the most psychologically attractive representation (Sperber 1996). Things get even more complicated when we consider that variations in the frequency of innovations may have a selecting effect: Innovations are relatively rare in a well functioning and stable society, whereas cultural crisis stimulates innovations. A dysfunctional culture, full of stress and conflicts, will have more innovations for selection to work on than a well functioning culture that makes everybody happy. The dysfunctional culture is therefore likely to have a shorter lifespan.

Those readers who want more exact mathematical formulations of certain distinct cultural selection mechanisms are referred to the literature reviewed in chapter 2, in particular Boyd and Richerson (1985), who also discuss the influence of genes on cultural selection criteria and mechanisms.


3.7 Selection of meme complexes

Some cultural selection processes are more aptly viewed as the selection of entire meme complexes rather than single memes. This applies for example to the spread of religious cults. New cult members buy the whole package of beliefs and lifestyle rather than just a single meme. Typical parts of a meme complex include:

This is a promised benefit that make the meme complex look attractive to potential new hosts. For example the promise of a better life. Often the bait involves button pushing to get the attention of the potential host. The bait may be a trojan horse.
This is the part of a meme complex that urges replication. In the case of religion this is typically a command to evangelize. In the case of a pyramid scheme business or chain letter the hook is an economical incentive to recruit new members.
Making sure that the host acquires all the memes in the meme complex. This may involve frequent repetition or brain washing. Rituals, songs, prayers, and oaths are typical examples.
Protection against rival meme complexes
This could be the meme that blind faith is a virtue and that heresy is a sin.
Reward and punishment
Obviously, rewards or punishments are often needed to make the host obey the instructions of the meme complex and its organization. A typical immediate reward is the belonging to a supportive social group. Sometimes the meme complex serves all the basic needs of the host. But the most powerful promised rewards and punishments are often due in a distant future or in the afterlife so that it is too late for the host to change his mind if the promises are not effectuated. Obvious examples are Hell and Paradise.
The demand that the host contribute time, energy, or money to the meme complex and its organization. These resources are needed by the organization for the purpose of competition against rival meme complexes.


3.8 Vicarious selection

Cultural selection processes can often be described as vicarious selection (Campbell 1965. Also called preselection). The principle behind vicarious selection is that a slow and ineffective selection process is supplanted by a faster and more effective selection process leading in approximately the same direction, whereby the adaptability is increased. The vicarious mechanism is in some way created by the old selection process, and may possibly be checked by the latter - albeit ineffectively - if it runs away. Campbell mentions our choice of food as an example. If we eat something inappropriate we may die from malnutrition or poisoning, so the choice of food is ultimately determined by natural selection. Our immediate choice, however, is based on taste. The genetic evolution has designed our sense of taste in such a way that healthy food tastes good. The taste criterion is approximately equal to the criterion of nourishment, and in this way the selection based on taste has become a replacement for the much slower selection based on survival.

Another example of vicarious selection is sexual selection. Many women prefer men that look strong and healthy. Physical strength is, at least in primitive societies, an important survival factor. The individual selection based on survival will promote physical strength, but the sexual selection based on partner choice preempts this process and weeds out weak individuals faster than the individual selection process can. In other words: sexual selection is vicarious for individual selection.

The best example of vicarious selection is rational choice. Our rules of hygiene, for example, have been introduced with the conscious aim of avoiding epidemic diseases. The same rules could in principle have arisen by the evolution of an instinctive urge to wash oneself. Or they could have arisen by cultural selection if religious cleanliness-taboos, innovated by chance, had been spread as a consequence of the fact that those who obeyed these taboos had lower risk of dying from infectious diseases than others. The intelligent choice of hygiene rules is vicarious for the much slower selection mechanisms based on survival or death. The advantage of vicarious selection is obvious: It means a faster and more effective adaptation to changing life conditions, and lower costs in terms of deaths. The vicarious selection mechanism will therefore in itself be adaptive, and therefore be promoted by the process for which it is vicarious (Boehm 1978).

There is, however, a serious complication in vicarious selection mechanisms, namely that they never lead in exactly the same direction as the processes they supplant. Let us return to the taste example. In our society the choice of food based on taste has lead to an over-consumption of sugar and to the addition of artificial flavorings to unhealthy foods. These possibilities have arisen so fast that the genetic selection has not had the time to check our taste. The most famous example of a vicarious selection process running away is the peacock's tail, which has evolved by sexual selection despite the fact that it reduces the viability of the peacock (see chapt. 4.2).

The concept of vicarious selection is important because the cultural selection process as a whole acts vicariously for the genetic selection, and indeed very effectively so. Many different selection mechanisms, including conscious choice, can be described as examples of vicarious selection.


3.9 Barriers in evolution

Biological evolution is not necessarily characterized by gradual change, as previously believed. Rather, evolutionary history may display saltatory periods of rapid change alternating with periods of relative quiescence, called punctuated equilibria (Somit & Peterson 1992). When a certain threshold has been passed, then the evolution goes relatively fast until a new equilibrium is established. The crossing of such thresholds has low probability and consequently happens very infrequently.

One possible explanation of such barriers in evolution is that the fitness effect of two different genes need not be additive. This phenomenon is called epistasis (Moore & Tonsor 1994). Assume, for example, that a hypothetical trait AB is coded by two dominant genes A and B on two independent loci. An individual that carries both these genes will have the phenotype AB which gives high fitness. An individual that has either A or B, but not both, will on the contrary have a lower fitness than an individual which has neither. It is totally unlikely that the genes A and B will arise by mutation simultaneously in the same individual. The evolution of AB must therefore necessarily pass through an intermediary stage involving hybrid individuals who have A or B, but not both. These hybrids will always be rare because they have lower fitness. It therefore requires a great deal of luck for a mating between an A individual and a B individual to take place and produce an AB offspring. Once this has happened, there is a reasonably high probability that the new trait will spread (under certain conditions), and that other genetic changes will follow to fine-tune the new adaptation, leading to a new punctuated equilibrium. If a hypothetical trait requires more than two genes to be combined before a fitness gain is seen, then the probability is virtually nil.

This type of probability barriers in genetic evolution is the reason why new organs rarely arise in evolutionary history. It is more probable that an existing organ will be modified to a new function than that a totally new organ be created from scratch, because the former process requires fewer simultaneous gene changes than the latter. This is, for example, the reason why a bird's wings are homologous with our arms.

Similar barriers are often seen in cultural evolution9. A new idea can not easily gain a footing in a society if it does not make sense or if it is not compatible with the existing rules and structure of that society (Burns & Dietz 1992). The impediment occurs when the utilization of a new idea only is possible when several memes in society are changed simultaneously and these changes have considerable costs. As an example of this, Rambo (1991:87) mentions the well-known problems of exporting high technology to developing countries. The most well studied kind of cultural barrier crossings is in the history of the sciences, where such a barrier crossing is called a paradigm shift (Kuhn 1962).

In genetic evolution, the probability of a barrier crossing is extremely small if it requires more than two genes to change simultaneously. In cultural evolution, however, such a probability problem can be overcome by intelligent planning. Just think of all the complicated technological inventions that cultural evolution has created. A modern technical appliance can contain thousands of components, and if just one single component is missing then the whole device will not work. A similar degree of complexity cannot possibly arise by natural evolution unless every intermediate step in the evolution has a slightly higher fitness than the previous one.

Let me give an example of such a compatibility barrier in our society. If we introduced a decimal system for measuring time, rather than hours, minutes, and seconds, then all calculations of time differences would be easier. Such a change of time units would however have considerable costs because all social institutions and technical devices are adapted to the old impractical units. Things get even worse when we consider that we have a coherent system of measuring units. The second is a basic unit from which other units like joule and volt are derived. Decimalization of our time units would therefore require that all derived units also had to be changed. The short-term costs of such a change of measuring units would be so immense, that we renounce the long-term advantage, however obvious it may be.

This type of barriers may be impossible to cross in genetic evolution, but in the cultural process they can be overcome by intelligent planning and investment if the long-term advantage is sufficiently attractive.

Cultural compatibility barriers are sometimes constructed more or less deliberately in order to manipulate the selecting action of others. Social actors possessing intellectual power can define a certain situation or a certain social problem by means of a certain frame of reference, and thereby determine which kind of actions are possible in relation to this problem. Burns & Dietz (1992) characterize this strategy as "redefining a situation to increase the likelihood of a desired outcome." More about this in chapter 8.


3.10 Differences between genetic and cultural selection

The formal similarities between genetic and cultural evolution models should not mislead anybody to draw conclusions by analogy from one model to the other. The differences between the two processes are so important and fundamental that the cultural selection model should be regarded as an independent theory. The cultural process is extraordinarily much faster and more effective than the genetic process, for several reasons which I will list here for clarity10:


3.11 Cultural selection in animals

Social scientists have traditionally viewed culture as something uniquely human. But if we define culture as a behavior pattern common to a group of individuals and which is transmitted from individual to individual by imitation or learning rather than by genetic inheritance, then there is no doubt that culture can be found in animals, and there are numerous documented examples of behavior patterns in animals which are transmitted by learning (Bonner 1980, Gardner 1994, Heyes & Galef 1996).

A touching example, which is always cited when the talk is about cultural transmission in animals, shall not be withheld here: In 1953, a one and a half year old japanese macaque named Imo found out that she could wash off the sand from sweet potatoes by rinsing them in water before eating them. After four and a half years, 18% of the adult monkeys and 79% of the juveniles in the troop had learnt the potato washing technique by imitating Imo. In 1961, all monkeys born later than 1950 had acquired the technique except one. Masao Kawai has studied the spreading pattern for the potato-washing behavior and documented a connection with the social structure. In 1956, Imo made another invention. She found out that she could separate wheat grains from sand by dropping them into the water so that the wheat grains would float and the sand grains would sink. This invention has spread in a similar way (Kawai 1965; Watanabe 1994).

The ability to learn from conspecifics is not limited to our closest relatives in the animal kingdom. Another famous example is about birds: Great tits have learnt to open the top of milk bottles and get access to the cream. This skill has arisen accidentally in a few places in Northern Europe, and from these centers of innovation the behavior has spread by imitation. The transmission by imitation is not limited to conspecifics. Related species like blue tits have been observed to imitate the bottle-opening behavior of the great tits (Fisher & Hinde 1949, 1951).

The song patterns of birds, the croaking of frogs, etc., are behaviors that young animals learn from their older conspecifics. This way of transmission leads to local dialects which may help the animals identify relatives from their home district (Mundinger 1980, Slater & Williams 1994).

Occasionally, it has been possible to document that the entire social organization of a group of animals has been modified as an adaptation to changed ecological conditions. Connie Anderson (1989) has observed a group of baboons which during a few years changed their social organization and mating pattern as a consequence of the presence of a predator. I will return to this example on page 139.


3.12 The applicability of cultural selection theory

As mentioned in chapter 2, cultural selection theory has been known since 1867. From my study of more than a hundred years' literature on this subject, it appears that the most striking problem with cultural selection theory has been that this theory rarely has been applied to the study of real world phenomena until recently. Those examples, which have been used to support the theory, have mainly been about rational selection, e.g. that new crops which give better yields are increasingly replacing older less effective crops. Such a conclusion is so trivial that an elaborate theory to explain it seems superfluous. In other words: Cultural selection theory has mostly been used to explain what was already obvious.

In my opinion, cultural selection theory has its greatest force in the area of irrational behavior. All societies are full of seemingly irrational and unproductive activities, such as religion, rituals, myths, tales, dance, music, festivity, art, fashion, play, sport, hobby, sex, and romance. All these activities have changed immensely during history, and we are seldom able to tell why. This is really a challenge for cultural selection theory.

But also rational decisions may have interesting selection effects. Egoistic decisions taken by influential persons or groups may have unwanted consequences for other groups or for the society as a whole. This leads us to conflict research, where selection theory also may be useful. If we can uncover the factors that determine the outcome of a conflict, then we may in principle be able to predict the macro level combined effect of a thousand micro level conflicts.

Functionalistic explanations in social theory have often been unconvincing for lack of a detailed etiological theory. In traditional analyses of the class struggle, for example, it has often been claimed that this or that institution exists "because it serves the interests of the ruling class". The traditional analysis often fails, however, when it tries to find the architect behind the strategy of the ruling class. A close scrutiny may often reveal that such a strategy may be more refined and artful than the members of the ruling class can possibly have been able to think out and agree upon. In particular, this may be the case when we are talking about religious, ideological, or other cultural means. Such strategies of power cannot possibly be explained by rational planning alone, but only by taking into account the accumulated effect of repeated selecting events.

In the following chapter, I will further develop the cultural selection theory in order to improve its explanatory power. In the rest of the book I will apply the theory to historical and contemporary cultural phenomena. I will focus mainly on the long term cumulative effect of many small selecting events, on conflicts of power and on irrational phenomena such as religion and art.

As already explained, it is impossible to formulate a universal mathematical model that will describe any cultural selection process. Both innovation mechanism, reproduction mechanism, selection mechanism, and unit of selection depend on the specific phenomenon under study. Most cultural phenomena are subjected to several different selection mechanisms which all interact with each other. An exact mathematical model for a real life cultural phenomenon like art or religion would be immensely complicated and have so many unknown parameters that it would be useless. I have therefore chosen to concentrate on qualitative rather than quantitative models in my study of cultural selection.


8. The word 'primal' is missing in the standard english translation, but the german original says 'Urvater'.

9. See Hodgson (1996) for the example of economic growth.

10. See also Daly (1982) and Ball (1984).

11. Some bacteria show an increased mutation rate under starvation. The mechanism behind this is still unresolved (Hall 1990, Sniegowski & Lenski 1995).