1 Un « fait frappant »… « The most striking fact about the relationship between evolutionary game theory and economic game theory is that, at the most.

11

Un « fait frappant »…Un « fait frappant »…

« The most striking fact about the relationship between evolutionary game theory and economic game theory is that, at the most basic level, a theory built of hyperrational actors and a theory built of possibly non-rational actors are in fundamental agreement. This fact has been widely noticed, and its importance can hardly be overestimated »

(Skyrms 2000, p. 273)

« The most striking fact about the relationship between evolutionary game theory and economic game theory is that, at the most basic level, a theory built of hyperrational actors and a theory built of possibly non-rational actors are in fundamental agreement. This fact has been widely noticed, and its importance can hardly be overestimated »

(Skyrms 2000, p. 273)

22

Rationality & EvolutionRationality & Evolution1 Two Optimizing Processes2 Isomorphism

2.1 Sex and Fairness2.1.1 Rationality as (economic) game theory

2.2 Evolution as evolutionary game theory2.2.1 Biological evolutionary game theory2.2.2 Economic evolutionary game theory

2.3 Evolutionary Generalism2.4 Why isomorphism is important

3 Differences3.1 Symmetry3.2 Dominated Strategies3.3 Actual Fitness and Intended Outcomes

4 Simulations and multi-level models4.1 Axelrod4.2 A methodology: Multi-level modeling4.3 Computational Differences

5 Normativity5.1 Normative Rationality vs. Descriptive Evolution5.2 What are the prospects for convergence?

1 Two Optimizing Processes2 Isomorphism







33















44

« All that the evolution-hypothesis tells us is that certain kinds of conduct are more evolved than others; […]. Yet [Mr. Spencer] tells us that one of the things he has proved is that conduct gains ethical sanction in proportion as it displays certain characteristics. What he has tried to prove is only that in proportion as it displays those characteristics, it is more evolved, it is plain, then, that Mr. Spencer identifies the gaining of ethical sanction with the being more evolved. »

Moore, Principia Ethica, pp. 48-4.

« All that the evolution-hypothesis tells us is that certain kinds of conduct are more evolved than others; […]. Yet [Mr. Spencer] tells us that one of the things he has proved is that conduct gains ethical sanction in proportion as it displays certain characteristics. What he has tried to prove is only that in proportion as it displays those characteristics, it is more evolved, it is plain, then, that Mr. Spencer identifies the gaining of ethical sanction with the being more evolved. »

Moore, Principia Ethica, pp. 48-4.

Le « sophisme naturaliste »Le « sophisme naturaliste »

55

L’heuristique de la personnalisation permet d’échapper au sophisme naturaliste…

L’heuristique de la personnalisation permet d’échapper au sophisme naturaliste…« If natural selection controls which

of traits T, A1, A2,…, An, evolves in a given population, then T will evolve, rather than the alternatives listed, if and only if a rational agent who wanted to maximize fitness would choose T over A1, A2,…, An »

(Sober 1998, pp. 408f.)

« Rational choice only gets value out by optimizing the value input captured by the agent’s preferences. Conversely, what evolution optimizes is only a value if reproductive fitness is valued. »

(p. 3)

« If natural selection controls which of traits T, A1, A2,…, An, evolves in a given population, then T will evolve, rather than the alternatives listed, if and only if a rational agent who wanted to maximize fitness would choose T over A1, A2,…, An »

(Sober 1998, pp. 408f.)

« Rational choice only gets value out by optimizing the value input captured by the agent’s preferences. Conversely, what evolution optimizes is only a value if reproductive fitness is valued. »

(p. 3)

66

Fitness individuelle vs. fitness collectiveFitness individuelle vs. fitness collective« And as natural selection works

solely by and for the good of each being, all corporeal and mental endowments will tend to progress towards perfection »

(Darwin 1859,p. 489)

« The economists of Darwin’s time tended to think that since a society is ‘nothing but’ a collection of individuals, the society will maximize its well-being if each individual endeavors to maximize his welfare. … »

(Sober 1984)

« And as natural selection works solely by and for the good of each being, all corporeal and mental endowments will tend to progress towards perfection »

(Darwin 1859,p. 489)

« The economists of Darwin’s time tended to think that since a society is ‘nothing but’ a collection of individuals, the society will maximize its well-being if each individual endeavors to maximize his welfare. … »

(Sober 1984)

77

Dépendance envers la fréquence:la tragédie du pré communal

Dépendance envers la fréquence:la tragédie du pré communal« [T]he rational herdsman concludes

that the only sensible course for him to pursue is to add another animal to his herd. And another.... But this is the conclusion reached by each and every rational herdsman sharing a commons. Therein is the tragedy. Each man is locked into a system that compels him to increase his herd without limit -- in a world that is limited. Ruin is the destination toward which all men rush, each pursuing his own best interest in a society that believes in the freedom of the commons. Freedom in a commons brings ruin to all. »

Garrett Hardin, Science, 162(1968):1243-1248.

« [T]he rational herdsman concludes that the only sensible course for him to pursue is to add another animal to his herd. And another.... But this is the conclusion reached by each and every rational herdsman sharing a commons. Therein is the tragedy. Each man is locked into a system that compels him to increase his herd without limit -- in a world that is limited. Ruin is the destination toward which all men rush, each pursuing his own best interest in a society that believes in the freedom of the commons. Freedom in a commons brings ruin to all. »

Garrett Hardin, Science, 162(1968):1243-1248.

88

By the way, c’est exactement ce qui est arrivé à nos morues…

By the way, c’est exactement ce qui est arrivé à nos morues…

99

L’évolution optimise quand même la rationnalité…

L’évolution optimise quand même la rationnalité…« While grazers’ and fishers’

welfare will not be maximized by evolution in a commons, their rationality should be. Rationality, after all, is the perfection of just those abilities useful for exploiting any situation, including social dilemmas. »

(p. 4)

« While grazers’ and fishers’ welfare will not be maximized by evolution in a commons, their rationality should be. Rationality, after all, is the perfection of just those abilities useful for exploiting any situation, including social dilemmas. »

(p. 4)

1010

Le « rationalisme évolutionniste »Le « rationalisme évolutionniste »« Creatures inveterately wrong

in their inductions have a pathetic but praise-worthy tendency to die before reproducing their kind. »

(Quine 1969)

« Natural selection guarantees that most of an organism's beliefs will be true, most of its strategies rational »

(Dennett 1987)

« Creatures inveterately wrong in their inductions have a pathetic but praise-worthy tendency to die before reproducing their kind. »

(Quine 1969)

« Natural selection guarantees that most of an organism's beliefs will be true, most of its strategies rational »

(Dennett 1987)

1111















1212

Le problème de l’égalité (numérique) des sexesLe problème de l’égalité (numérique) des sexes« [G]enerally females are the scarcity

constraint on reproduction, so we should expect fewer males in an optimum sex mix. »

(p. 5)

« I formerly thought that when a tendency to produce the two sexes in equal numbers was advantageous to the species, it would follow from natural selection, but I now see that the whole problem is so intricate that it is safer to leave its solution for the future »

(Darwin 1871)

« [G]enerally females are the scarcity constraint on reproduction, so we should expect fewer males in an optimum sex mix. »

(p. 5)

« I formerly thought that when a tendency to produce the two sexes in equal numbers was advantageous to the species, it would follow from natural selection, but I now see that the whole problem is so intricate that it is safer to leave its solution for the future »

(Darwin 1871)

1313

1414

Problème parallèle : séparer le gâteauProblème parallèle : séparer le gâteau• On sait que, pour être juste, il

faut séparer le gâteau 50-50, mais la théorie des jeux ne nous dit pas pourquoi.

• Équilibre de Nash:« We have an equilibrium in

informed rational self-interest if each of our claims are optimal given the other’s claim. In other words, given my claim you could not do better by changing yours and given your claim I could do no better by changing mine. »

(Skyrm 1996, p. 5)

• On sait que, pour être juste, il faut séparer le gâteau 50-50, mais la théorie des jeux ne nous dit pas pourquoi.

• Équilibre de Nash:« We have an equilibrium in

informed rational self-interest if each of our claims are optimal given the other’s claim. In other words, given my claim you could not do better by changing yours and given your claim I could do no better by changing mine. »

(Skyrm 1996, p. 5)

1515

Problème: il y a une infinité d’équilibres de Nash

Problème: il y a une infinité d’équilibres de Nash

000

11

11

Dem

ande

AD

eman

de A

Demande BDemande B

1616

Est-ce que c’est moi qui ne comprend pas?Est-ce que c’est moi qui ne comprend pas?« Skyrms argues that rational choice

cannot answer this basic question about fairness »

(p. 6)

« Skyrms argues that rational choice cannot answer this basic question about fairness »

(p. 6)

Stratégie (demande x)Stratégie (demande x)

Util

ité a

ttend

ueU

tilité

atte

ndue

1717

Payoffs par stratégie, en termes de reproductive fitness

Payoffs par stratégie, en termes de reproductive fitness

Demande 1/3

Demande 1/2

Demande 2/3

ModesteDemand

e 1/31/3 1/3 1/3

Impartial

Demande 1/2

1/2 1/2 0

CupideDemand

e 2/32/3 0 0

1818

Fitness attendues, avec distribution égale des stratégies

Fitness attendues, avec distribution égale des stratégies

Demande 1/3

Demande 1/2

Demande 2/3

Fitness attendue

ModesteDemand

e 1/31/3 1/3 1/3 1/3

Impartial

Demande 1/2

1/2 1/2 0 1/3

CupideDemand

e 2/32/3 0 0 2/9

1919

Fitness attendues, une fois que les cupides sont disparus

Fitness attendues, une fois que les cupides sont disparus

Demande 1/3

Demande 1/2

Demande 2/3

Fitness attendue

ModesteDemand

e 1/31/3 1/3 1/3

Impartial

Demande 1/2

1/2 1/2 1/2

CupideDemand

e 2/3

SESSES

2020

Une version en automate cellulaireUne version en automate cellulairehttp://www.ags.uci.edu/~jalex/lattice-models/http://www.ags.uci.edu/~jalex/lattice-models/

2121

Dynamique de l’égalitéDynamique de l’égalité

Demande 1/3Demande 1/3 Demande 2/3Demande 2/3

Demande 1/2Demande 1/2

2222

C’est la même chose pour le sex ratioC’est la même chose pour le sex ratio

(Fisher 1930, p. 142)(Fisher 1930, p. 142)

2323















2424

Rationality as (economic) game theoryRationality as (economic) game theory« Evolution is only isomorphic

to rationality if we restrict the range of both concepts. » (p. 8)

• Danielson choisi de se limiter à la rationalité dans les interactions, qui est le problème le plus complexe de la rationalité.

« Evolution is only isomorphic to rationality if we restrict the range of both concepts. » (p. 8)

• Danielson choisi de se limiter à la rationalité dans les interactions, qui est le problème le plus complexe de la rationalité.

2525















2626

La théorie des jeux s’applique mieux à la biologie qu’à l’économie…

La théorie des jeux s’applique mieux à la biologie qu’à l’économie…« There are two reasons for this.

– First, the theory requires that the values of different outcomes […] be measured on a single scale. In human application, this measure is provided by ‘utility’ – a somewhat artificial and uncomfortable concept: in biology, Darwinian fitness provides a natural and genuinely one-dimensional scale […].

– Secondly, and more importantly, in seeking the solution of a game, the concept of human rationality is replaced by that of evolutionarily stability. The advantage here is that there are good theoretical reasons to expect populations to evolve to stable states, whereas there are grounds for doubting whether human beings always behave rationally. »

(Maynard Smith 1982, p. vii).

« There are two reasons for this.– First, the theory requires that the values of

different outcomes […] be measured on a single scale. In human application, this measure is provided by ‘utility’ – a somewhat artificial and uncomfortable concept: in biology, Darwinian fitness provides a natural and genuinely one-dimensional scale […].

– Secondly, and more importantly, in seeking the solution of a game, the concept of human rationality is replaced by that of evolutionarily stability. The advantage here is that there are good theoretical reasons to expect populations to evolve to stable states, whereas there are grounds for doubting whether human beings always behave rationally. »

(Maynard Smith 1982, p. vii).

2727















2828

Qu’est-ce qu’une SES?Qu’est-ce qu’une SES?

« It is a strategy such that, if most of the members of a population adopt it, there is no ‘mutant’ strategy that would give higher reproductive fitness »

(Maynard Smith & Price 1973, p. 15)

« For distinct strategies x and y and utility function u,

1) u(x,x) ≥ u(y,x)2) If u(x,x) = u(y,x) then u(x,y) >

u(y,y) »(p. 10)

« It is a strategy such that, if most of the members of a population adopt it, there is no ‘mutant’ strategy that would give higher reproductive fitness »

(Maynard Smith & Price 1973, p. 15)

« For distinct strategies x and y and utility function u,

1) u(x,x) ≥ u(y,x)2) If u(x,x) = u(y,x) then u(x,y) >

u(y,y) »(p. 10)

2929















3030

Economic evolutionary game theoryEconomic evolutionary game theory« Where biological

evolutionary game theory is intentionally broad in the scope of its agents, economic evolutionary game theory focuses more narrowly on explaining human action. [E]conomic evolutionary game theory modeling is based on a human learning dynamic. »

(p. 11-12)

« Where biological evolutionary game theory is intentionally broad in the scope of its agents, economic evolutionary game theory focuses more narrowly on explaining human action. [E]conomic evolutionary game theory modeling is based on a human learning dynamic. »

(p. 11-12)

3131















3232

Un pas de plus: généralisme évolutionniste

Un pas de plus: généralisme évolutionniste« [Fair division’s] strong stability

properties guarantee that is an attracting equilibrium in the replicator dynamics, but also make the details of the dynamics unimportant. Fair division will be stable in any dynamics with a tendency to increase the proportion (or probability) of strategies with greater payoffs … For this reason, the Darwinian story can be transposed into the context of cultural evolution, in which imitation and learning may play an important role in the dynamics »

(Skyrms 1996 , p. 11)

« [Fair division’s] strong stability properties guarantee that is an attracting equilibrium in the replicator dynamics, but also make the details of the dynamics unimportant. Fair division will be stable in any dynamics with a tendency to increase the proportion (or probability) of strategies with greater payoffs … For this reason, the Darwinian story can be transposed into the context of cultural evolution, in which imitation and learning may play an important role in the dynamics »

(Skyrms 1996 , p. 11)

3333















3434

Retour sur la citation de départRetour sur la citation de départ« The most striking fact about the

relationship between evolutionary game theory and economic game theory is that, at the most basic level, a theory built of hyper-rational actors and a theory built of possibly non-rational actors are in fundamental agreement. This fact has been widely noticed, and its importance can hardly be overestimated. Criticism of game theory based on the failure of rationality assumptions must be reconsidered from the viewpoint of adaptive processes. There are many roads to the Nash equilibrium concept, only one of which is based on highly idealized rationality assumptions »

(Skyrms 2000, p. 273)

« The most striking fact about the relationship between evolutionary game theory and economic game theory is that, at the most basic level, a theory built of hyper-rational actors and a theory built of possibly non-rational actors are in fundamental agreement. This fact has been widely noticed, and its importance can hardly be overestimated. Criticism of game theory based on the failure of rationality assumptions must be reconsidered from the viewpoint of adaptive processes. There are many roads to the Nash equilibrium concept, only one of which is based on highly idealized rationality assumptions »

(Skyrms 2000, p. 273)

3535















3636

SymétrieSymétrie

« A single population evolutionary setting imposes a symmetry requirement which selects Nash equilibria which appear implausible in other settings »

(Skyrms 2000, p. 273)

« [Evolution] often (but not always) leads to selection of fair division in a simple bargaining game. »

(Skyrms 1996, ch. 1)

« A single population evolutionary setting imposes a symmetry requirement which selects Nash equilibria which appear implausible in other settings »

(Skyrms 2000, p. 273)

« [Evolution] often (but not always) leads to selection of fair division in a simple bargaining game. »

(Skyrms 1996, ch. 1)

3737















3838

Deux différences: stratégies faiblement dominées et rationalité modulaire

Deux différences: stratégies faiblement dominées et rationalité modulaire• “[R]efinements of the Nash

equilibrium are handled differently.– Standard evolutionary

dynamics […] does not guarantee elimination of weakly dominated strategies.

– [E]volutionary dynamics need not eliminate strategies which fail the test of sequential rationality »

(Skyrms 2000, p. 273)

• “[R]efinements of the Nash equilibrium are handled differently.– Standard evolutionary

dynamics […] does not guarantee elimination of weakly dominated strategies.

– [E]volutionary dynamics need not eliminate strategies which fail the test of sequential rationality »

(Skyrms 2000, p. 273)

3939

Concept de dominanceConcept de dominance• When a player tries to choose the "best"

strategy among a multitude of options, that player may compare two strategies A and B to see which one is better.– B dominates A: choosing B always gives at least

as good an outcome as choosing A. There are 2 possibilities:

• B strictly dominates A: choosing B always gives a better outcome than choosing A, no matter what the other player(s) do.

• B weakly dominates A: There is at least one set of opponents' action for which B is superior, and all other sets of opponents' actions give A and B the same payoff.

• This notion can be generalized beyond the comparison of two strategies.– Strategy B is strictly dominant if strategy B

strictly dominates every other possible strategy.– Strategy B is weakly dominant if strategy B

dominates all other strategies, but some are only weakly dominated.

http://en.wikipedia.org/wiki/Dominance_(game_theory)

• When a player tries to choose the "best" strategy among a multitude of options, that player may compare two strategies A and B to see which one is better.– B dominates A: choosing B always gives at least

as good an outcome as choosing A. There are 2 possibilities:

• B strictly dominates A: choosing B always gives a better outcome than choosing A, no matter what the other player(s) do.

• B weakly dominates A: There is at least one set of opponents' action for which B is superior, and all other sets of opponents' actions give A and B the same payoff.

• This notion can be generalized beyond the comparison of two strategies.– Strategy B is strictly dominant if strategy B

strictly dominates every other possible strategy.– Strategy B is weakly dominant if strategy B

dominates all other strategies, but some are only weakly dominated.

http://en.wikipedia.org/wiki/Dominance_(game_theory)

4040

Rationalité modulaireRationalité modulaire« In a credible contingency plan for a situation

in which an agent faces a sequence of choices, her plan should specify a rational choice at each choice point, relative to her situation at that choice point »

(Skyrm 1996, p. 24)

« In a credible contingency plan for a situation in which an agent faces a sequence of choices, her plan should specify a rational choice at each choice point, relative to her situation at that choice point »

(Skyrm 1996, p. 24)

4141

Le jeu de l’ultimatumLe jeu de l’ultimatum

Si joueur 1

Si joueur 2

Gamesman

S1

Offre 1 Accepte tout

S2

Offre 1 Rejette tout

S3

Offre 1 Accepte 5, rejette 1

Mad Dog S4

Offre 1 Rejette 5, accepte 1

Easy Rider

S5

Offre 5 Accepte tout

S6

Offre 5 Rejette tout

Fairman S7


S8


4242

Le jeu de l’ultimatum:répartition égale des stratégies

Le jeu de l’ultimatum:répartition égale des stratégies

Si joueur 1

Si joueur 2 % init.

Gamesman

S1

Offre 1 Accepte tout 12.5

S2

Offre 1 Rejette tout 12.5

S3


12.5

Mad Dog S4


12.5

Easy Rider

S5

Offre 5 Accepte tout 12.5

S6

Offre 5 Rejette tout 12.5

Fairman S7


12.5

S8


12.5

4343

Le jeu de l’ultimatum:les Mad Dogs (faiblement dominés) survivent!

Le jeu de l’ultimatum:les Mad Dogs (faiblement dominés) survivent!

Si joueur 1

Si joueur 2 % init.

% final

Gamesman

S1

Offre 1 Accepte tout 12.5 87

S2

Offre 1 Rejette tout 12.5 0

S3


12.5 0

Mad Dog S4


12.5 13

Easy Rider

S5

Offre 5 Accepte tout 12.5 0

S6

Offre 5 Rejette tout 12.5 0

Fairman S7


12.5 0

S8


12.5 0

4444

En version automate cellulaireEn version automate cellulairehttp://www.ags.uci.edu/~jalex/lattice-models/http://www.ags.uci.edu/~jalex/lattice-models/

4545

Le jeu de l’ultimatum:haute proportion initiale de fairmen

Le jeu de l’ultimatum:haute proportion initiale de fairmen

Si joueur 1

Si joueur 2 % init.

Gamesman

S1

Offre 1 Accepte tout 32

S2

Offre 1 Rejette tout 2

S3


10

Mad Dog S4


2

Easy Rider

S5

Offre 5 Accepte tout 10

S6

Offre 5 Rejette tout 2

Fairman S7


40

S8


2

4646

Le jeu de l’ultimatum:les fairmen (faiblement dominés) survivent

Le jeu de l’ultimatum:les fairmen (faiblement dominés) survivent

Si joueur 1

Si joueur 2 % init.

% final

Gamesman

S1

Offre 1 Accepte tout 32 0

S2

Offre 1 Rejette tout 2 0

S3


10 0

Mad Dog S4


2 0

Easy Rider

S5

Offre 5 Accepte tout 10 43.5

S6

Offre 5 Rejette tout 2 0

Fairman S7


40 56.5

S8


2 0

4747

Dynamique de l’ultimatumDynamique de l’ultimatum

4848















4949

« Rational choice is concerned with the intended outcomes of action. Selection mechanisms operate through actual outcomes. In explanations of animal behavior, where intentions have at best a minimal place, actual outcomes must bear most of the explanatory burden. It is more controversial which mechanism is the most important in the study of human action »

(Elster 1989, p. 71)

« Rational choice is concerned with the intended outcomes of action. Selection mechanisms operate through actual outcomes. In explanations of animal behavior, where intentions have at best a minimal place, actual outcomes must bear most of the explanatory burden. It is more controversial which mechanism is the most important in the study of human action »

(Elster 1989, p. 71)

5050















5151

Méthodes: simulations vs. modèlesMéthodes: simulations vs. modèles« Roughly, the contrast

between rationality and evolution projects onto methods, with evolution characterized more by simulations and rationality by more formal models. »

(p. 16)

« Roughly, the contrast between rationality and evolution projects onto methods, with evolution characterized more by simulations and rationality by more formal models. »

(p. 16)

5252















5353

AxelrodAxelrod• Axelrod 1984: The Evolution of

Cooperation– Évolution des stratégies dans le dilemme du

prisonnier itéré;– Stratégie gagnante: TIT-FOR-TAT

• Critiques:– On connaissait déjà le folk theorem…– Trop d’emphase sur TIT-FOR-TAT;

• Binmore 1998:« The folk theorem of game theory proved by

several authors simultaneously in the early fifties … describes in precise detail all of the outcomes of a repeated game that can be sustained as equilibria. [Axelrod] did us the service of focusing our attention on the importance of evolution in selecting an equilibrium from the infinitude of possibilities whose existence is demonstrated by the folk theorem. »

• Axelrod 1984: The Evolution of Cooperation– Évolution des stratégies dans le dilemme du

prisonnier itéré;– Stratégie gagnante: TIT-FOR-TAT

• Critiques:– On connaissait déjà le folk theorem…– Trop d’emphase sur TIT-FOR-TAT;

• Binmore 1998:« The folk theorem of game theory proved by

several authors simultaneously in the early fifties … describes in precise detail all of the outcomes of a repeated game that can be sustained as equilibria. [Axelrod] did us the service of focusing our attention on the importance of evolution in selecting an equilibrium from the infinitude of possibilities whose existence is demonstrated by the folk theorem. »

5454















5555

Évolution et rationalitéÉvolution et rationalité

« Stepping back, it is obvious that the complete picture must include both; evolution has historically produced some rational agents. So the question arises, how might one model the interaction between evolution and rationality? Given the difference in time scales of the two processes, a natural way to approach the relation of rationality and evolution is by a two level model.»

(p. 18)

« Stepping back, it is obvious that the complete picture must include both; evolution has historically produced some rational agents. So the question arises, how might one model the interaction between evolution and rationality? Given the difference in time scales of the two processes, a natural way to approach the relation of rationality and evolution is by a two level model.»

(p. 18)

5656

Exemple de modélisation multi-niveauExemple de modélisation multi-niveau

5757















5858

Mieux que rationnel?Mieux que rationnel?« One point is particularly important for economists

to appreciate: it can be demonstrated that ‘rational’ decision-making methods … are computationally very weak: incapable of solving the natural adaptive problems our ancestors had to solve reliably in order to reproduce… This poor performance on most natural problems is the primary reason why problem-solving specializations were favored by natural selection over general purpose problem-solvers. …. On evolutionarily recurrent computational tasks, such as object recognition, grammar acquisition, or speech comprehension the human mind greatly outperforms the best artificial problem-solving systems that decades of research have produced, and it solves large classes of problems that even now no human-engineered system can solve at all. »

(Cosmides et Tooby 1994)

« One point is particularly important for economists to appreciate: it can be demonstrated that ‘rational’ decision-making methods … are computationally very weak: incapable of solving the natural adaptive problems our ancestors had to solve reliably in order to reproduce… This poor performance on most natural problems is the primary reason why problem-solving specializations were favored by natural selection over general purpose problem-solvers. …. On evolutionarily recurrent computational tasks, such as object recognition, grammar acquisition, or speech comprehension the human mind greatly outperforms the best artificial problem-solving systems that decades of research have produced, and it solves large classes of problems that even now no human-engineered system can solve at all. »

(Cosmides et Tooby 1994)

5959















6060

Pourquoi la rationalité est normative, mais pas l’évolution?

Pourquoi la rationalité est normative, mais pas l’évolution?• Comparez:

« Pourquoi devrais-je me préoccuper d’être copié? »

« Pourquoi devrais-je me préoccuper de mes propres préférences? »

• Comparez:« Pourquoi devrais-je me

préoccuper d’être copié? »« Pourquoi devrais-je me

préoccuper de mes propres préférences? »

6161

Les paradoxes de la rationalitéLes paradoxes de la rationalité• La rationalité est auto-

référentielle: il est rationnel d’être rationnel.– Notre concept de rationalité

est fondé sur un « mélange d’intuition, d’analogies et d’idéologie » (Samuelson 1997)

– L’évolution peut peut-être• Modifier ces intuitions• Nous éclairer sur leur source

• La rationalité est auto-référentielle: il est rationnel d’être rationnel.– Notre concept de rationalité

est fondé sur un « mélange d’intuition, d’analogies et d’idéologie » (Samuelson 1997)

– L’évolution peut peut-être• Modifier ces intuitions• Nous éclairer sur leur source

6262

L’influence du positif sur le normatifL’influence du positif sur le normatif« If I have strong reasons, based on

evolutionary equilibrium selection, to expect you to choose a particular strategy, normative rationality gives me a reason to coordinate with your choice. »

(p. 22)

« If possibility is construed generously we have utopian theory. Those who would deal with ‘men as they are’ need to work with a more restrictive sense of possibility. »

(Skyrms 1996, pp. 108-9)

« If I have strong reasons, based on evolutionary equilibrium selection, to expect you to choose a particular strategy, normative rationality gives me a reason to coordinate with your choice. »

(p. 22)

« If possibility is construed generously we have utopian theory. Those who would deal with ‘men as they are’ need to work with a more restrictive sense of possibility. »

(Skyrms 1996, pp. 108-9)

6363















6464

Chacun tire la couverte…Chacun tire la couverte…« [T]he combination of education and evolution

drives society in the direction of [game] theory. …. A widely applicable theory of games would, of necessity, involve a strong element of self-prophecy in the sense that the existence of the theory itself would be partly responsible for bringing about stabilizing the event which it ‘predicts’ »

(Binmore 1990, p. 18f).« One might then reasonably expect to see

[constrained choice in cooperative dilemmas] drive out more costly precommitment and enforcement methods, and this through nothing more than what economists like to describe as the ordinary competitive process. »

(McLennen 1998)

« [T]he combination of education and evolution drives society in the direction of [game] theory. …. A widely applicable theory of games would, of necessity, involve a strong element of self-prophecy in the sense that the existence of the theory itself would be partly responsible for bringing about stabilizing the event which it ‘predicts’ »

(Binmore 1990, p. 18f).« One might then reasonably expect to see

[constrained choice in cooperative dilemmas] drive out more costly precommitment and enforcement methods, and this through nothing more than what economists like to describe as the ordinary competitive process. »

(McLennen 1998)

6565

Fin.Fin.

1 Un « fait frappant »… « The most striking fact about the relationship between evolutionary game theory and economic game theory is that, at the most.

Documents

economic game theory

rationality evolution

evolutionary generalism

descriptive evolution

multilevel models

multilevel modeling

normative rationality

computational differences