Summary: the evolved social communication style of humans is not optimised for problem-solving. People of average intelligence generally fail to notice this, therefore they habitually fail to make full use of their faculties of precise thinking, and privilege conversation at the expense of less natural but effective means of problem-solving such as books, the internet and humans outside their social group.
Dunbar’s number is a theoretical upper limit to the number of other people with whom an average human can maintain stable social relationships, given the limitations of human brain power in facilitating these relationships. The number is extrapolated from a regression of mean group size in other primates and the neocortex volume of these primates; the human average neocortex volume is input, and out pops Dunbar’s number which is given to be ~150 (although the error bars are rather wide).
Other primates, with their relatively small social group sizes, maintain their social relationships by physically grooming one another. However, the large potential social group size of humans renders physical grooming impractical because it would be so time consuming.
From the wikipedia article on Dunbar’s number:
Dunbar, in Grooming, Gossip, and the Evolution of Language, proposes furthermore that language may have arisen as a “cheap” means of social grooming, allowing early humans to efficiently maintain social cohesion. Without language, Dunbar speculates, humans would have to expend nearly half their time on social grooming, which would have made productive, cooperative effort nearly impossible. Language may have allowed societies to remain cohesive, while reducing the need for physical and social intimacy.
This certainly seems to be the case – humans use conversation to maintain their social relationships, rather than picking insects and dead skin off each other (if only we could claim to be so much more advanced than our primate cousins in every respect).
The relevance of this to rationality, as I see it, is that to the extent that human conversation is not optimised for efficient exchange of information and problem solving, social conversation is a sub-optimal means of obtaining help from other people. Presumably, our natural mode of conversation with social companions is a compromise between conversation as instrumental means of solving problems or gaining information, and conversation as grooming – it seems to me that in the EEA, these criteria would not have led to entirely the same results if our approach to conversation were only optimised for one or the other.
If there had been internet and books in the EEA, humans would have evolved to prefer these information sources to what they can extract from their social contacts, in comparison to their natural inclinations given our actual evolutionary history (in which conversation was usually the best information source available, compromise though it may be). Given this counterfactual I suppose conversation would be optimised more as a purely social grooming activity, too.
Therefore when people are talking to friends, they may be unduly optimistic about how useful they expect these interactions to be in comparison to (for example) researching things themselves, or asking questions on an internet forum. Admittedly humans are generally well aware that ulterior motives can be at play in their conversations, but even if someone has good reason to believe that their interlocutor is not consciously trying to hinder them (as is generally the case with friends) they are likely to overestimate the quality of the information, analysis or advice they receive.
To state the obvious, social companions may be unwilling to speak the truth, regardless of how important it is to the problem. More subtly, friends may fail to realise the degree of precision of thought that is necessary in solving certain problems and resolving dilemmas – they may think they are offering advice (and the person with the problem may think so too), when in fact they are still in the habitual mode of conversation optimised towards the standards of “be interesting”, “don’t make a status play by acting too smart or talking too long” et cetera.
I see this as a problem that relatively smart people are likely to circumvent, firstly because they are used to thinking precisely in their work or studies and know what that feels like, therefore they tend to recognise when their social interactions are ill suited to problem-solving; and secondly because “acting too smart” is not frowned upon as an attempt to claim undue status in their social groups. That’s probably why I’ve never seen this problem discussed on Less Wrong; nonetheless to the extent that anyone is interested in improving the practical rationality of the IQ<115 crowd, it might be a useful concept to tackle with them.
Conversation as social grooming – a rationality problem?
Summary: the evolved social communication style of humans is not optimised for problem-solving. People of average intelligence generally fail to notice this, therefore they habitually fail to make full use of their faculties of precise thinking, and privilege conversation at the expense of less natural but effective means of problem-solving such as books, the internet and humans outside their social group.
Dunbar’s number is a theoretical upper limit to the number of other people with whom an average human can maintain stable social relationships, given the limitations of human brain power in facilitating these relationships. The number is extrapolated from a regression of mean group size in other primates and the neocortex volume of these primates; the human average neocortex volume is input, and out pops Dunbar’s number which is given to be ~150 (although the error bars are rather wide).
Other primates, with their relatively small social group sizes, maintain their social relationships by physically grooming one another. However, the large potential social group size of humans renders physical grooming impractical because it would be so time consuming.
From the wikipedia article on Dunbar’s number:
This certainly seems to be the case – humans use conversation to maintain their social relationships, rather than picking insects and dead skin off each other (if only we could claim to be so much more advanced than our primate cousins in every respect).
The relevance of this to rationality, as I see it, is that to the extent that human conversation is not optimised for efficient exchange of information and problem solving, social conversation is a sub-optimal means of obtaining help from other people. Presumably, our natural mode of conversation with social companions is a compromise between conversation as instrumental means of solving problems or gaining information, and conversation as grooming – it seems to me that in the EEA, these criteria would not have led to entirely the same results if our approach to conversation were only optimised for one or the other.
If there had been internet and books in the EEA, humans would have evolved to prefer these information sources to what they can extract from their social contacts, in comparison to their natural inclinations given our actual evolutionary history (in which conversation was usually the best information source available, compromise though it may be). Given this counterfactual I suppose conversation would be optimised more as a purely social grooming activity, too.
Therefore when people are talking to friends, they may be unduly optimistic about how useful they expect these interactions to be in comparison to (for example) researching things themselves, or asking questions on an internet forum. Admittedly humans are generally well aware that ulterior motives can be at play in their conversations, but even if someone has good reason to believe that their interlocutor is not consciously trying to hinder them (as is generally the case with friends) they are likely to overestimate the quality of the information, analysis or advice they receive.
To state the obvious, social companions may be unwilling to speak the truth, regardless of how important it is to the problem. More subtly, friends may fail to realise the degree of precision of thought that is necessary in solving certain problems and resolving dilemmas – they may think they are offering advice (and the person with the problem may think so too), when in fact they are still in the habitual mode of conversation optimised towards the standards of “be interesting”, “don’t make a status play by acting too smart or talking too long” et cetera.
I see this as a problem that relatively smart people are likely to circumvent, firstly because they are used to thinking precisely in their work or studies and know what that feels like, therefore they tend to recognise when their social interactions are ill suited to problem-solving; and secondly because “acting too smart” is not frowned upon as an attempt to claim undue status in their social groups. That’s probably why I’ve never seen this problem discussed on Less Wrong; nonetheless to the extent that anyone is interested in improving the practical rationality of the IQ<115 crowd, it might be a useful concept to tackle with them.