Frozen Evolution. Or, that’s not the way it is, Mr. Darwin. A Farewell to Selfish Gene.

Some memes can spread very effectively even though they are disadvantageous for their bearers. The meme for smoking spreads in spite of the fact that smoking demonstrably shortens life expectancy and worsens the health of its bearer and persons living in his vicinity, i.e. most frequently his biological relatives, and thus reduces his inclusive fitness (Kunzle et al. 2003; Munafo et al. 2002). The success of the meme for smoking is not only a result of its physiological addiction and the fact that it is imitated by adolescents as a symbol or maturity (and the maintenance of the smoking habit in adulthood is then ensured by the already-mentioned addiction). Smoking, similar to the consumption of chocolate or hard drugs, is pleasant for the individual, at least initially or at the time of consumption. The memes that an individual will attempt to adopt are decided, not by the degree to which they increase or decrease his fitness but by the degree to which they increase his feeling of pleasure or reduce feelings of stress (see XV.2). Even such obviously disadvantageous behavior as suicide, or behavior disadvantageous for its bearer but advantageous for his surroundings, such as some patterns of altruistic behavior, tends to be imitated. Because cultural evolution occurs incomparably faster than biological evolution, there is very little hope that selection against genes that are biologically disadvantageous, i.e. genes determining that biologically disadvantageous behavior will be perceived as unpleasant, could make a species immune to spreading of the particular disadvantageous memes. For example, it cannot be expected that the bad habit of smoking or over-eating sweets would, in time, be reduced by natural selection in that the individuals that find these bad habits pleasant would gradually disappear from the population because they would produce fewer offspring on an average.
            In general, the same rules apply to the spreading of memes as to the spreading of infectious diseases and the relevant processes can also be described by the same formal models (Anderson 1993). The most important parameter that determines the efficiency with which a meme will spread is its basic reproduction constant R0, a dimensionless constant equal to the average number of individuals “infected” by the relevant meme by one bearer of the particular meme in the “naive” population, i.e. a population whose members had not previously encountered the meme. If this constant is larger than 1, the given gene can spread in the population and be retained in it for a long time, even if it is harmful to its bearers, i.e. if it reduces their fitness. The actual reproduction constant, R, in a population in which a certain fraction of individuals, q, was “infected” by the meme in the past, is understandably lower and decreases linearly with increasing fraction of infected individuals. If R decreases to a value of 1, the fraction of infected individuals remains constant in the population and the particular meme is retained endemically in the population. Each meme is differently “infectious” and each has a different threshold intensity, NT, i.e. number of susceptible individuals, at which it can begin to spread in the population. A simple equation exists between the threshold density and R0 
 
      (1)
 
where N is the size of the population. If “meme infection” has occurred in the population, the value of R0 can be calculated from the fraction of individuals that remained unaffected by the meme (s), as it holds that
 
     (2)
 
On the basis of R0 we can then easily calculate the size of the fraction of persons in the population that it would be required to make immune to the particular meme through an effective campaign so that this meme would not be able to spread by horizontal transmission.
 
   (3)
 
If, for example, a certain meme affects an average of 70% of individuals during its natural horizontal spreading in the population, then it would be necessary to “immunize” 41% of the so-far unaffected population in advance to prevent a future meme epidemic. Because the fraction of immunized persons gradually decreases in a natural way after the end of the epidemic, either through the deaths of immune individuals or forgetting, some meme epidemics can have a regular cyclic character, where the periodicity of the cycle will depend on the size of the population.
            Some other interesting laws, which have already been described for classical epidemiology, also govern the spreading of memes (Ewald 1994), see also the chapter XIX.5. Memes (similar to parasites) transmitted exclusively or predominantly vertically, i.e. from parents to children, generally do not harm their bearers, as their successful spreading is closely connected with the fitness of their bearers. In contrast, memes that can spread horizontally can be far more harmful for their bearers. This is especially true of memes that are not transmitted horizontally by direct personal contact between neighbors, but tend to be transmitted over long distances, in the modern world, for example, through the press and television, and are simultaneously not bound to a particular culture, so that they can spread transculturally. Especially harmful memes can spread in populations whose members have a reduced life expectancy for some reason, for example as a result of a war, poor nutrition or diseases. The positive feedback effect can also be important here, where the spreading of the harmful meme reduces the life expectancy of the members of the population, enabling effective spreading of even more harmful memes. The gradual reduction in the occurrence of all possible forms of individual or mass aggression during the second half of the 20th century can be most readily explained as a side effect of the prolonged life expectancy of human beings. The longer this life expectancy, either as a consequence of improved hygiene or advances in medicine, the greater are the penalties for memes that might be successful in the short run, but harm their bearers in the longer term.
The harmfulness of memes is further increased by the possibility of “superinfection”, i.e. increased probability of simultaneous infection of a single person by several memes. If this possibility is negligibly small, an advantage is usually given to those memes that allow their hosts to live as long as possible, so that they are capable of “infecting” a large number of other individuals in the population. However if, during infection by one meme, there is a danger of infection by another meme, even memes that harm their bearers very rapidly can be successful, for example the meme for use of hard drugs. Mutability of memes acts similarly to superinfection. From this point of view, for example, religious systems based on canonized texts will probably be less harmful to their adherents than the religious systems of various sects. Population growth is a factor that can promote the spreading of dangerous memes; memes that are beneficial for their bearers tend to spread in populations with stable sizes or those that are diminishing in size.s