Socialization: Mores and Emotions

Transmission of information can occur both genetically and culturally.  Although the former is often more widespread than the latter (taste for sweet v. taste for bread), both can be “unlearned.”  In addition, norms and mores can persist even when they diminish fitness (pre-industrial cities, smoking, small families, eating brains of deceased relatives).

One of the results of cultural transmission is the transformation of norms and mores into emotive preferences, so that “oughts” are internalized and become “wants.”  This cannot be achieved by

·         punishment because it produces fear and aversion, not positive wants.

·         general intelligence because some sort of emotional attachment is necessary, as psychopaths show.

 

Cultural transmission is vertical (from parents) or oblique (from elders, schools, etc), and often based on conformism, the innate tendency of young humans to imitate behavior, or on institutions whose goal is to socialize the young.

Two related puzzles:

1.       Why did the capacity to internalize norms evolve, since the capacity to internalize is costly?

2.       Why did individually costly norms become internalized, thus producing costly preferences?

The answer (B&G chs. 10-11) follows a pleiotropic analogy: the capacity to internalize norms and the associated creation of costly wants are tied to the internalization and creation of advantageous preferences. The following models by B&G support that answer. 

 

Analytic model for costly internalization gene

One genetic locus for capacity to internalize, which is the expression of one gene with two alleles a and b: a (capacity to internalize) is costly; b (no such capacity) is not.  (Allele a is just a placeholder for the internalization phenotypic trait, as almost certainly more than one gene is involved).

·         Genetics are haploid, but random mating is diploid with usual Mendelian segregation.

·         Normless phenotype D has baseline fitness 1; norm C, “Be clean!”, for example, is beneficial conferring fitness 1+f>1.  Note that only a-individuals can be C-individuals.  An a-offspring is equally likely to be C or D.

·         Allele a costs u, while allele b costs 0.

·         Fitness is multiplicative, so that the fitness of aC is (1-u)(1+f)>1. (This makes the math easier).

·         Frequency of phenogenotype is proportional to fitness.

·         Only vertical transmission takes place.  The strength of vertical cultural transmission of C is β= 0…,1: P(a-offspring of CC is C)=1; P(a-offspring of DD is D)=1; ceteris paribus, P(a-offspring of CD is C)=1/2.  Note that the strength of vertical transmission of D is 1- β, which means that D has vertical transmission as well.

 

Results

Three equilibriums of homogeneous populations:

1.       All aC: Locally stable

2.       All aD:  Unstable

3.       All bD: Locally stable.  Becomes unstable when β>1/2, and this results in aC becoming globally stable.

As β>1/2 when C-parents are more effective than D-parents at transmitting values, for plausible degrees of cultural vertical transmission, the capacity to internalize norms becomes fixated.

Diploid model, studied by Gintis, has relevantly similar results.

 

Analytical model for costly internalized norm as hitchhikers.

Same as previous model plus:

·         Internalizable norm A which is group promulgated but individually costs 0<s<1, while normless state S costs 0; only a-individuals can internalize A.

·         In addition to vertical transmission there is oblique transmission.  a-individuals who are S-types adopt A with probability pγ, where p is the frequency of A-types, γ= 0…, 1 is the strength of oblique transmission of A. (There’s never oblique transmission of S, as A is socially preferred).  Hence, the gain in A-types from this mechanism is pγ(1-p).

 

Results

There are 5 equilibriums, all with homogeneous populations:

1.       aAC  Stable if s < γ, which occurs when the strength of oblique transmission overcomes the cost of A.  

2.       aAD  Unstable

3.       aSC Stable if s > γ

4.       aSD  Unstable

5.       bSD Stable if (1-u)(1+f) < 2, which is likely

So, the fitness advantage of C renders the internalization allele a viable.  As a moves towards fixation, if the strength of oblique transmission is sufficiently larger than the fitness cost of A, then A will become fixated.

 

Analytical model with payoff-based updating

Same model a before plus

·         Every a-individual compares the strategy of a randomly chosen individual; if the other’s is better, he switches with probability 0<μ <1 times the difference in fitness of the two types. (μ measures the imitation rate).  So, some aCA could switch to aCS, for example.

Results

1.       All aAC is stable if

2.       All aSC is stable if , so that aSC is stable even if s < γ.

Hence, when payoff updating is allowed, γ must be significantly larger than s to obtain the condition for aAC’s stability.

So, the fitness advantage of C renders the internalization allele a viable.  As allele a moves towards fixation, if the strength of oblique transmission is sufficiently larger than the fitness cost of A, then A will become fixated.  However, if payoff updating is allowed, then the strength of oblique transmission must be significantly larger than A’s cost. 

So if aSC’s achieves stability within the group, weak group selection is enough to spread aAC’s in the population as long as A is prosocial. 

 

Agent based model for internalization of altruistic norms

 

·         1000 groups, 36 individuals each in 3 generations of 12 per group.  Groups on a 50x50 torus.

·         Start: 2% aAC; 1% aAD; 1% aSC; 1% aSD; 95% bSD

·         S= .03; f= .06; u= .01;  benefit of altruism to group such that on average members of all A-types group is are 5% more fit than those of all S-types group.

·         Each group has randomly assigned μ and γ.  Random variations in social arrangement make them increase or decrease by 1% of their value.

·         Migration rate is 25% per generation, always to neighboring groups.  No intentional segmentation.  When group falls below 4, it’s replaced by copy of most successful neighboring group.

·         Mutation rate .01% per generation

·         Conflict rate 10% per generation with a randomly chosen neighboring group; group with higher fitness prevails, and defeated group copies winning group.

·         Costs of institutions for oblique transmission to A-types:

o   γ has a cost of γs for each A-type

o   combating μ, the value of the lure of higher payoff in payoff updating, costs s(1- μ) to each A-type 

·         Updates are fitness based for each generation in each group.

 

Result:

System stabilizes rapidly (about 200 generations) to either no A’s or to a high frequency of A’s (between 50% and 75% depending on the cost for A’s). The latter always occurs with at least an initial 2% of A’s.  High frequency of A’s also results in low μ (about 25%) and γ (about .08).  Aside from the frequency of initial A’s, the values of initial parameters are not critical.  

So, there are plausible mechanisms resulting in internalization of prosocial norms. 

The costs and benefits of internalization are due to the following:

Costs:

·         Cognitive: the part of the brain that has to be devoted to this

·         Social: resources spent in socializing younger generation

·         Diminished flexibility: some internalized rules, parochial ones, for example, reduce fitness-increasing opportunities.

Benefits:

·         Avoidance of ego-depleting calculation.  This seems very widespread in humans

·         Avoidance of disastrous mistakes, at least most of the times.  (Mill’s rules of thumb)

·         Possibility of overriding individual fitness maximization in favor of group-fitness increase.

 

So, weak or strong group selection is crucial for the success of the internalization of social norms and associated emotions.  Consequently, although group selection underdetermines the specific content of such norms/emotions (they could be consequentialist, deontological, or whatever), they must be prosocial.   

But why did we become so different from Pan, given that we started with a recent common ancestor about 6mya?

Boehm hypothesis: group hunting large ungulates may have been a turning point in creating or accelerating the appearance of cultural routines that

1.       Averted serious conflict in meat eating

2.       Guaranteed that every member of the hunting group would get enough food to remain an efficient hunter.  (This became very significant in the late Pleistocene when short range climate variation was intense).

(1)-(2) favored the emergence upside-down dominance and the success of individuals who exercised self-restraint; this favored those gene pools that curbed excessive aggression or cunning, and this genetic change, in turn, favored the appearance of norm internalization and the beginning of moral conscience.