© Werner Leinfellner
Evolution is a game - the greatest game
in the world if you know how to play it.
I. A SHORT HISTORY OF EVOLUTIONARY GAME THEORY
Since Neumann and Morgenstem's theory of games, the debate among social scientists, economists, mathematicians, and social philosophers about what kind of theory it is has not ended. Some think that it is a new interdiscipline, some that it is a mere accumulation of gametheoretical models, such as utility theory, competitive, cooperative, collective choice models, and so on. Most of them agree that the models of game theory deal with isolated, single, and independent specific societal interactions between individuals who wish to maximize their gains and minimize their losses within prescribed constraining rules.
The situation changed abruptly when game theorists found out that most games depend on previous games, on past experiences, on their history which rest on a specifically human, not always rational way of thinking. There is no doubt that we leam from our past successful and unsuccessful attempts, from iterating games, from improving and adapting rules, and exploiting random events. Here the simelarity between series of causally interdependent competitive, cooperative games, serial games against nature, and evolutionary processes, is striking. The same applies when we leam to adapt losses or gains to changing rules and to selective random events in our society and environment. In the eighties, these similarities led to the new name "evolutionary game theory. " More and more astonishing similarities with evolutionary processes were detected, for example: Evolutionary games follow numerous societal background rules; they are causally dependent on previously played games and on their stored rules; serial winnings resemble the measurable "fitness-increase" - but in a non-Darwinian sense. Series of temporally ordered value matrices were used to explain and successfully compute the dynamics of evolutior)ary serial games in animal populations. It is no wonder that since the eighties evolutionary game theory began to merge with Neodarwinism, with the evolution equations of modern biological evolution theory, of molecular biology, and of population genetics into a new unified theory of evolution. This new theory attempts to explain all kinds of evolutionary processes. Its methods and models in fact cover the prebiotic, molecular, and biological evolution of organisms and the evolution of animal and human behavior in their societies and environments.
When this new theory of evolutionary dynamics was applied to societal evolution, many inconsistencies within (the game-theoretical model of) the predominant Darwinian traditional evolution theory emerged. The leading role of Darwinian competition was replaced by the self-organization of societal cooperation Evolutionary game theory reached a peak when it succeeded in modelling and reformulating Darwinian selection and adaptation. Today, a new, twofold, mathematical-stochastic method explains the evolutionary sociodynainics of societies. One component is the dynamic evolutionary matrix method developed from evolutionary game theory. The other uses differential equations, called "evolution equations" . Both, when applied to societal and cultural evolution, explain in a statistico-causal sense why societal forms change from Generation to Generation; moreover, they explain the creation of new and the disappearance of old societal forms within prescribed, rulelike constraints. Kauffman's grand view that all evolutionary processes exist only "near the edge of chaos" entails both: that randomness subdues societal cultural evolution and that it enables the self-organization of a more and more complex democratic order. It also explains the ävolution of complex molecules in prebiotic evolution. Or, as genetic algorithm, it is, according to Holland, a new method to explain genetic evolution.
This article tries to apply the unified theory of evolution to societal and cultural evolution in modem, democratic societies and tries to remove some hindrances and obstacles, a task already tackled by Lumbsden & Wilson, Maynard Smith, and Helbing, to mention just a few.
11. THE FIRST OBSTACLE: Do WE PLAY GAMES OR DO WE SOLVE
CONFLICTS IN OUR PRESENT DEMOCRATIC SOCIETIES?
A PARADIGM CHANGE.
It makes no empirical sense to speak about playing games with and in our present society. According to Selten, most games of game theory are only abstract, mathematical models of how we may compute their outcomes. Like so many abstract concepts, the application of gametheoretical models to explain societal evolution runs into almost insurmountable difficulties. The most serious is: All evolutionary processes rest on causal or statistico-causal interactions between the individuals and between the individuals and their environment. Till today, game theory offered no viable empirical 'microscopic' foundation or interpretation of these interactions. This article proposes a new empirical interpretation of serial, evolutionary games, which explains how games can be interpreted as societal, empirical interactions. Solving games in mente or solving them theoretically does not change anything in our societal life. But when we realize our solutions of societal conflicts, we do interact empirically with our society. Only our empirical realizations cause changes in our sociocultural evolution and in our environment. Empirical realizations of our mental or theoretical solutions of evolutionary conflicts are the real causes which push forward socio-cultural evolution.
T'his amounts to a paradigm change in today's evolutionary game theory from ,playing games" and "gambling" to "solving and realizing serial conflicts" in our society. To solve a problem is, according to H. Simon, a theoretical effort; to solve a societal conflict includes its empirical realizations. Moreover, empirical solutions need additional, democratically accepted background rules which we impose on the few special gametheoretical rules, as in Harsanyi's social ethicsor in Holland's performance system. Not the gametheoretical special rules, but the societal, juridical, cultural, and socio-ethical empirical rules for the realizations of theoretically possible solutions play the most important role. One began to doubt that the traditional, rational methods can be used when we realize empirically evaluations in utility theory under uncertainty and risk. Allais' paradox and most of the paradoxes which have haunted and shaken utility and game theory and related disciplines, including postmodem constructivism and many cognitive sciences, suffer from the reality syndrome. It indicates that, in reality or when we realize our solutions of societal conflicts, everything looks suddenly paradoxical and different than in the abstract theory. We may solve these paradoxes through a preventive measure: Not the abstract theory, only its empirical realization counts.
III. THE COMPLEXITY INCREASE OF THE SOCIETAL CULTURAL EVOLUTION
Since the complexity of the socio-cultural evolution increased beyond the complexity of the biological and animalic evolution, the complexity of the corresponding theories increased. Firstly, the dynamic evolutionary mechanism of the theory of socio-cultural evolution explains the societal and cultural changes by, firstly, an additional information- and rule-transfer to future generations by individual and cultural storage, learning and teaching. This can only be achieved optimally in modern democracies. Secondly, by the paradigmatic change of one-shot games to serial solutions and empirical realizations of societal conflicts; all individual preferences, evaluations, and decisions of how to solve societal conflicts can be taken into account, for example, by collective choice theory, a typical democratic theory. Thirdly, we have an ever-increasing imposition of new and stored socio-ethical background rules on gametheoretical rules in democracies. Fourthly, by virtue of the fact that our brain's intelligent faculties to model and solve societal conflicts consciously and scientifically, the brain functions as a real 'prime mover' of societal and cultural evolution. Fifthly, mental, scientific, and even creative conflict-solving is not enough: Only those mental or theoretical solutions 'survive' which tum out to be the product of practically, technically, and empirically successful processes of realizations. These realizations are the long-sought empirical interactions and at least partial causes of the socio-ethical evolution; at the same time, they serve as their empirical tests. Only realizations may become man-made, differently weighted, partial causes for successful or unsuccessful evolutionary changes in our sociocultural settings. Sixthly, according to Weibull and Leinfellner, evolutionary game theory permits not only the explanation of the macrodynamics of societal evolution in the large, but also of the evolutionary mini-dynamics of how our intelligent brain solves mentally societal conflicts. When we - mentally or scientifically - solve societal conflicts or problems, our methods and thought processes are to be evolutionary mini-methods, too, or small-scale mental evolutionary processes. They differ from the traditional, merely rational, linear and recursive methods by their specific stochastic inferences and by the default rules.
For example, when we leam, when we invent and create better solutions, better empirical realizations, we have to cope with imagined randomness. When we solve societal conflicts, the mini-evolutionary methods of thinking rest in most instances not on rational deductive methods, not on classical statistical methods, but on non-linear, partly stochastic Bayesian methods under uncertainty and risk and on statistico-causal inferences and beliefs. For example, when we adapt our solutions of societal conflicts to random changes in our environment, when we empirically interpret mixed strategies, or when we cooperate, the empirical solutions drift randomly. Since evolutionary, stochastic thinking does not use exclusively rational deductive methods, it resembles - according to Simon, Leinfellner, and many brain physiologists statistico-causal inferencing. Besides the rational, predominantly axiomatic and deductive, thinking, reserved for first-hand shaping of models, this way of evolutionary cognitive and practical thinking is our second, inborn, 'deeper' intelligent capacity for gathering and evaluating social experience under uncertainty and risk with the help of numerous causal and prescriptive default rules which are stored in our individual and cultural memory. Another reason is: Here we use statistical inferences to solve and realize our pending, future societal conflicts; and we do this only for the better of each individual's and the entire society's welfare. This describes the main attractor in the evolution of democratic societies.
The next important problem for democratic societies is the socio-ethical complexity of its stored democratic rules. For example, when we solve societal conflicts, we obey not only a vast number of democratic, juridical, social, and ethical rules, but we also follow, at the same time, very often hidden evolutionary attractors, non-teleological tendencies, such as certain overriding socioethical rules. Only they can increase our own individual well-being and, at the same time, society's welfare and security in a Pareto-optimal sense.
IV. SOCIETAL CONFLICTS AND CONFLICT SOLVING
Do we really gamble when we solve societal conflicts? Apparently not, because, in contrast to private conflicts, societal conflicts are always open conflicts between the individuals' egoistic and altruistic interests within their society. "Egoistic" means here regarding only the selfish interests, "altruistic" regarding also the interests of the others. Open societal conflicts are solved theoretically by 'optimal' solutions. Such solutions are in most cases weighted, evolutionarily stable equilibria between egoistic and altruistic, extreme solutions. This holds equally for all societal conflicts copying with random events which effect some or all members of the society, such as contagious diseases, natural catastrophes, etc. Therefore, optimal or theoretically possible Nash equilibria are only necessary conditions for optimal solutions of societal conflicts. There is no other way: Empirical realizations within democratic societies have to increase
individual as well as collective well-being and welfare - not "in the long run"
but in the "not too long tun." In other words: They have to satisfy the Pareto
principle: Optimal are solutions whose realizations increase, or at least do not
decrease, all the individuals' and the society's welfare and stability. By no means can some individuals or groups in democracies of the social-welfare type profit at the expense of the other individuals or groups.
Even the best abstract framework and its mathematical solutions will not help us to solve our daily societal conflicts and problems, if their realizations violate democratic human rights or democratic rules which are based on the Pareto principle. Therefore, very often "theoretically optimal, rational, and consistent" solutions of societal conflicts, proposed by governments or specialists, are voted down by the majority of citizens who follow their democratically accepted rules.
In practice, serial realizations of societal solutions are a temporal, complex, Bayesian, evolutionary leaming process, since this process unfolds always by including new empirical evidence. When we have to decide between gametheoretically sound, but multiple, equilibria (possible solutions), the useftilness of their empirical realization, not rational deductive arguments, are decisive for a society. And only democracies dispose of enough democratically accepted socio-ethical rules which follow the Pareto principle. They pernüt only such empirical applications of societal conflict solutions as do not violate human rights and Bayesian learning.
VI. PROBLEMS OF COEVOLUTION
What most Neodarwinians do not want to admit is that societai and cultural evolution is an empirically quite different kind of evolution than biological evolution, the reason being that we ourselves effect the empirical realizations of societal solutions, including those of today's technical, medical and, hopeftilly, environmental conflicts. This is one reason why we cannot reduce the evolution of societal behavior in human societies to biological genetic evolution. Another reason is that coevolution is a holistic, synergetic, complex evolutionary flow which cannot be split up into components. Its evolutionary dynamic rests not only on mutually linked interactions, but can be steered by our active realizations of societal conflict solutions, at least in part. Our scientific, technological, our biological, medical, and socio-cultural know-how permits us to interfere with our socio-cultural evolution to a far greater extent than, for example, with our basic genetic and biological evolution. We may change the former by the realizations of competitive and cooperative conflict solutions; we may change the causal impact of selective random events from our environments and from intemal changes; and, of course, we may alter our adaptive rules for socio-cultural evolution - see example .
Therefore, research in societal-cultural evolution focusses today on the empirical, the psychological, the brainphysiological, and the intelligent partial causes of the socio-cultural coevolution. Recently, Damasio has shown how the realizations of societal solutions under selective pressure and adaptational rules depend on additional brainphysiological emotional causes.
VII. THE SOCIO-CULTURAL EVOLUTION IS AN EVOLUTION
OF SOCIETAL FORMS
As already described, a theory or the models of our societal and cultural evolution have to explain, firstly, how individuals in democratic societies solve mentally and scientifically all their societal conflicts. Secondly, theoretical solutions have to be realized; this is their ultimate empirical test. Iterations of societal solutions across generations have created an ever-increasing diversity of cultural forms. Cultural forms are mentifacts, such as ideas, models, theories; artifacts are artistic creations and creations by artisans; technifacts are products of our technological culture or civilization. Finally, we have pattems of behavior and of social activities, expressed by prescriptive rules and customs as cultural forms. Their respective evolutionary trajectories represent in the evolutionary models their temporal changes, their 'growth or increase' or 'extinction,' from Generation to Generation. The sum total of all changes in each evolutionary trajectory represents the changes in our socio-cultural evolution. Luckily, till today they increased steadily the societal welfare and stability of democratic societies.
VIII. REALIZATIONS OF SOCIETAL CONFLICTS ARE STATISTICO-CAUSAL
INTERACTIONS BETWEEN THE INDIVIDUALS AND THEIR ENVIRONMENT
lt is a fact that any kind of evolution rests on empirical causal interactions among the individuals, organisms and among them and their environments. However, to this day, even in gametheoretical evolutionary models, it was impossible to explain, according to Helbing, which 'microscopic' empirical causal interactions push forward the societal and cultural evolution. Our proposal to regard societal interactions as the empirical realizations of theoretically possible solutions of serial societal conflicts offers for the first time an empirical interpretation of all the man-made causes of socio-cultural evolution. For the author, all sciences, including the technical and the social sciences, serve only as a means to successfully realize our mentifacts, models, predictions, and technifacts empirically. Successful terminations of societal conflicts depend, on the one hand, on the conflict-solving constructive and inferencing mental activities of the individuals, and on the other hand, on their successful or unsuccessful realizations. This highlights the paradigm change from "solving games" to "solving conflicts and realizing the solutions". Therefore, societal and cultural evolution rests increasingly on individuals' readiness to solve mentally or with the help of gainetheoretical models any societal conflicts, as they arise, and on the practical empirical realizations of the theoretical solutions. The successful realization of solutions serves now as the ultimate empirical criterion, even for theoretically equivalent optimal solutions, for example, when we have to choose between optimal equilibria.
Carefully executed empirical realizations will prevent most of the haunting paradoxes which plague today's utility and game theory. Sociodynamics explains not only why and how temporally ordered sequences or strings of optimal societal solutions and their successful realizations change, but also why new societal forms originate, why they multiply, replicate, improve or deteriorate, or disappear. But let's keep in mind: Each realization of a societal conflict solution is only a partial cause among others which may influence causally the channeling or start of a societal form's evolutionary trajectory. So far we have had luck, according to Gould, since no chaotic changes of preferences or disturbing random events have impeded the rise of democratic societies.
Finally:
Since the successful empirical realizations of theoretical solutions of societal conflicts can change the socio-cultural evolution by creating or terminating societal forms, artifacts, mentifacts, technifacts, customs, and rules, the individuals are no longer passive onlookers of their socio-cultural evolution. They can change actively, at will, with the help of societal, political, economic, scientific and technical rules, the future course of their cultural evolution when they realize the solutions of societal conflicts. But in spite of these advantages, some individuals play a relatively passive role when their conflict solutions and realizations depend on the decisions or the free will of majority voters, on undemocratic power groups, or, in extreme cases on bad, catastrophic random events.
Thus, by cultural evolution we understand any change in any cultural forms or goods; it plays no role whether they are partially caused by the realizations of the solutions of conflicts and/or by selective random events.
The following models and their interpretations describe theoreticaly and empirically the causal connections which effect the specific changes of cultural forms or goods, and why they increase or decrease across generations.
But if we find, through intelligent and mental efforts and with the help of evolutionary models, theoretically possible optimal solutions, this is only a necessary and formal condition before we realize them. This first stage can be achieved by gametheoretical methods and/or stochastic reasoning. But without realizations, theoretically optimal solutions remain platonistic, mere mathematical constructions, as Selten has emphasized again and again. Only their confirmation by their empirical realizations - the relizations must fulfill the Pareto principle - can avert disastrous paradoxes or consequences. lt is as in all realizing sciences: New technifacts have to be invented mentally or theoretically; but without their empirical realization and proof of their usefulness they remain mentifacts, products of our constructive fantasy. The cultural self-organization of new goods has multicausal origins. It often arises from an unpredictable causal mix between intended, man-made realizations, provided positive random events, such as windfall profits in our economic system, support the creation of societal goods. See Bak's sandpile paradigm .
Cultural forms, by the way, come in two categories: (I) cultural, mental, and material goods; and (II) cultural, rule-governed pattems (activities). Cultural goods and cultural patterns are the inputs and outputs of our cultural evolution. Our realized solutions may give the ongoing cultural evolution a new direction, that is, channel new evolutionary paths; but they can shape only our immediate societal future. All evolutionary processes are openended. Thus we may interfere partially with the evolutionary macrodynamics of our culture by influencing causally certain evolutionary paths (trajectories). This may, with luck, streamline or channel for a while the sum total of all temporal changes of all mentifacts, artifacts, technifacts and societal pattems of behavior. Such a change can be expressed by default rules. Most of the time we are involved with channelling the microdynamics of single evolutionary trajectories of mentifacts, artifacts, behaviors in small units or subcultures. Evolutionary changes show up the first time as observable redistributions within subcultures before they Aggregate in larger societies. Presently, all empirical tests can check only the nücroevolutionary trajectories in subcultures, as described in the next section, not the total macrocultural evolution. Here exists approximately the same gap as between micro- and macroeconomics.