Most people can solve this familiar puzzle if they think about it for a while, but only slightly more complex versions have them floundering. Yet the problem described isn’t much more difficult than naming the day after the day after yesterday, which (I think) most people can do instantly. The fact that such a simple problem can be posed as a puzzle is just one piece of evidence that people (at least people in modern/Western societies) have trouble learning about and reasoning about kinship relations.Brothers and sisters I have none, but that man’s father is my father’s son
I’m generally sympathetic to the Cosmides-Tooby idea of the mind as a collection of special-purpose gadgets rather than a general-purpose computer. The work of Kahneman and Tversky on probability judgements (also my own main area of theoretical research) supports this idea. And I’ve even put forward evolutionary arguments to support the view that people are likely to overweight low-probability extreme events[1].
So, there is a bigger puzzle here for me. Assuming that the set of gadgets with which our minds are now equipped is the product of evolution, shouldn’t we (at least in some phase of our lives) be as good at learning about kinship systems as young children are at learning languages? After all, it’s hard to imagine that we can be acting to promote the survival of our genes if we don’t know who is carrying them.
It’s often asserted that modern/Western society has a particularly minimal kinship system and that the systems prevailing in other societies are considerably more complex. This certainly seems to be true of the Aboriginal Australian systems I’ve seen described, but I don’t know whether it’s true more generally. Has the kinship instinct atrophied over time, and, if so, what are the implications?
fn1. Contrary to the prevailing assumption that CTers are remorselessly opposed to evolutionary psychology in all its forms.
{ 15 comments }
John Isbell 04.09.04 at 11:19 pm
IIRC, Latin distinguishes maternal and paternal uncles (socer and avunculus), but for some reason it conflates grandson and nephew (nepos).
Peter Cuthbertson 04.10.04 at 12:23 am
The surest examples of inherent understanding and learning are those things we don’t know consciously. In your example, babies don’t pick up language through rational explanation because it is pure instinct. Certainly there’s no reason a natural, genetic disposition to be more charitable and kind to those to whom we are genetically close has to display itself at the rational level.
To take an example of which Richard Dawkins is fond, the shell of a snail conforms to logarithmic patterns that you’d really need a slide-rule or log table to produce. Does a snail have these things or understand logatharithms? Nope. It doesn’t need to.
It wouldn’t be proof that snails have no genetic capacity to produce their shells in the way they do that it is impossible to teach a snail logarithms, so there is no reason a comparable lack of facility for the puzzles you describe should prove kin relationships aren’t what they appear.
Peter Cuthbertson 04.10.04 at 12:39 am
“The embryological development of any bit of an animal’s or plant’s body requires complicated mathematics for its complete description, but this does not mean that the animal or plant must itself be a clever mathematician! Within any one species, the taller the tree, the relatively larger the buttresses. It is widely accepted that the shape and size of these buttresses are close to the economic optimum for keeping the tree erect, although an engineer would require quite sophisticated mathematics to demonstrate this. It would never occur to Sahlins or anyone else to doubt the theory underlying buttresses simply on the grounds that the trees lack the mathematical expertise to do the calculations. Why, then, raise the problem for the special case of kin selected behaviour? It can’t be because it is behaviour as opposed to anatomy, because there are plenty of other examples of behaviour (other than kin-selected behaviour, I mean) that Sahlins would cheerfully accept without raising his ‘epistemological’ objection; think, for instance, of my own illustration (p. 96) of the complicated calculations that in some sense we all must do whenever we catch a ball. One cannot help wondering: are there social scientists who are quite happy with the theory of natural selection generally but who, for quite extraneous reasons that may have roots in the history of their subject, desperately want to find something – anything – wrong with the theory of kin selection specifically?” – Richard Dawkins, The Selfish Gene (2nd Edition), p. 292
Steve Carr 04.10.04 at 1:05 am
I’m unconvinced that the failure to solve this riddle has anything to do with Westerners’ inability to reason about kin relations. (If this is one piece of evidence, I’d like to see the rest.) At the very least, what makes the riddle different from your example of the day-after-the-day problem is that the speaker in the riddle is externalizing his relationship to himself, speaking about himself as if he were someone else. That’s not the way any real person (at least in the West) ever talks, with the possible exception of pro athletes who refer to themselves in the third person. I think that’s the core of the riddle, not some putative difficulty in thinking about familial relationships.
John Quiggin 04.10.04 at 1:53 am
Peter, I’m not claiming that people should be able to give a formal and explicit account of how they know things. Obviously people can catch balls, identify and produce grammatical sentences and so on without any formal model. In the context of the puzzle and other questions about kinship, the appropriate analogy would be that people could easily identify the relevant relationship, but not explain how they did it.
It’s my observation that most people can’t do this in relation to anything but the simplest kin relationships. For example, what relation is my uncle to my niece? I think he’s a great-uncle but I don’t know for sure. Similarly I don’t know which of the two is closer to me genetically.
I suppose it’s possible to hypothesize that even though I don’t know these things, I may instinctively recognise the closer one and act to favor them, but this seems pretty desperate.
Finally, I should say that contra Dawkins, I’m quite happy about the theory of kin selection, since it obviously provides a genetic basis for altruism, which can then be extended by cultural generalisation. I am asking the question seriously, not as a debating point.
Peter Cuthbertson 04.10.04 at 2:32 am
There is no reason for people to recognise on any conscious level their closeness to another person, to the extent that the puzzle you describe is easy. Indeed, the evidence is that rationality operates on a different plane altogether: learning calculus and physics isn’t likely to make you better at catching a ball.
To take your example, your niece and uncle are both equally closely related, assuming your uncle is a full- and not half-brother of your mother or father. The chance of any particular gene in you being in a sibling or your offspring is in both cases 1/2 more than normal. The mathematics of your sibling’s child, or your parent’s sibling, having the same gene is therefore just 1/2 x 1/2. So your relatedness to both is 1/4 more than the normal for the species. Not difficult mathematics. Yet you think it is desperate to believe its principles could be recognised on anything but the conscious level, even as you presumably accept without question that every time someone catches a ball they have performed calculus without needing any conscious knowledge of the field. Why?
Barry 04.10.04 at 3:53 am
John Quiggin:
“…is just one piece of evidence that people (at least people in modern/Western societies) have trouble learning about and reasoning about kinship relations.”
No – at best it’s a piece of evidence that people in modern/Western societies don’t have the same amount of skill that people in more family-based societies do. That is, if there was some data which showed that people in such societies solved such puzzles more quickly. The obvious explanation, of course, is that people in less family-based societies spend less time thinking about familial relations.
Tom T. 04.10.04 at 6:13 am
John Q, I think part of the problem with evolutionary psychology is that it’s so easy to toss out contrary hypotheses. It seems to me that one could just as easily theorize that a kinship sense is not evolutionarily valuable, since genetic overlap drops so precipitously outside one’s immediate family.
Moreover, suppose we theorize that the preferred social grouping among early humans was a small, insular tribe of semi-nomadic hunter-gatherers. This means that most of your life would be spent among the same family members, many of whom may be related to you in multiple ways. Infusions of new genes would probably come from taking in young men driven out of other tribes, or capturing young women in war — in either case, if that person is your parent, there’s no evolutionary benefit to a kinship sense that includes their distant relatives, because their distant relatives are not around. Add to all this the fact a pre-verbal proto-human has no easy way of communicating to another proto-human, “Groo is your second cousin, Ugga is your niece, and Bob’s your uncle.” All in all, it seems plausible that our kinship sense would only develop to be binary — recognizing that someone else is either “one of us” or “not one of us.”
Keith 04.10.04 at 4:35 pm
I’d say that in our current society, memes are more important than genes so passing along our ideas takes precedence over who gets great Grandpa’s receeding hairline. This is part of the ongoing restructuring of the idea of family into a broader notion that encompasses friends, mates and former mates into the larger network of the Urban Tribe, which for many, fulfills the function of the traditional family. On a very pragmatic level, this is more useful to people than being able to site their lineage ten generations back.
msg 04.11.04 at 7:04 am
barry-
I agree with “…don’t have the same amount of skill that people in more family-based societies do…”
but I don’t think it’s from spending less time thinking about familial relationships.
It’s deeper, it’s identity. People don’t take their identity from the family anymore, in “modern” society.
There’s a real narrow band of social net that begins in the rigid age-based clans of school, that abruptly drops away. And is then replaced in a much more anxiety-fraught way by work clads and less importantly, small social groups. Family is vital to a lot of people still, but their identity doesn’t come from there.
LizardBreath 04.12.04 at 3:58 pm
This is a matter for empirical research, but anectdotally, I would believe that this:
“No – at best it’s a piece of evidence that people in modern/Western societies don’t have the same amount of skill that people in more family-based societies do. That is, if there was some data which showed that people in such societies solved such puzzles more quickly. The obvious explanation, of course, is that people in less family-based societies spend less time thinking about familial relations.”
is the answer you’re looking for.
I spent two years living in Western Samoa (Peace Corps). The standard conversation you have there, when you meet someone, is to try and figure out what connections you have (in the unusual case that you don’t already know). In a small country where families tend to be large, you are usually at no more than two degrees of separation with any other person. These conversations are all of the form of relationship puzzles: “So, Uili’s aunt’s husband, the one who came from Samatau, he was the one who was the pastor in Aopo?” “That’s right — his daughter married my cousin.” I think any puzzle like the one you posted about would be so immediately transparent to a Samoan used to this type of conversation that they wouldn’t even perceive it as a puzzle. Westerners simply don’t spend any time thinking about these sorts of relationship puzzles, and so have wildly underdeveloped skills.
John Quiggin 04.12.04 at 10:58 pm
Thanks for this, lizardbreath. This seems plausible to me, though problematic for the Cosmides-Tooby model.
dsquared 04.13.04 at 6:30 pm
Obliquely reminds me that I wanted to write something on the subject of why people, IMO, don’t “overweight” improbable tail events …
dsquared 04.13.04 at 6:39 pm
btw, John, how does the T’n’C model cope with the observable facts of head trauma and strokes?
If I understand things correctly, people who have strokes or brain injuries often suffer impairments of cognitive functions (proving that there is some mapping between neuroanatomy and cognitive function), but can almost always recover these functions with careful therapy (proving that the mapping of cognitive properties onto neuroanatomy is not set in stone).
Since the only way in which genetic inheritance could account for modularity of cognitive function would be through neuroanatomy, we either have to say that a brain-damage survivor who has relearned how to walk, feed himself and respond to visual stimuli does so in a way which is completely different from the way in which you or I do, or that there is no particular reason to believe in modularity of the T’n’C sort (that is, genetically determined modularity, rather than the fixing of learned behaviours in the neuroanatomy of the brain).
John Quiggin 04.13.04 at 10:55 pm
daniel, I don’t think the fact of post-stroke adaptation is fatal to the C-T model, but it points up its limitations. Even if the brain isn’t ‘designed’ as a general-purpose computer it is powerful, flexible and adaptable enough to serve as one, particularly when aided by culture.
A striking illustration is the way we’ve managed to exploit our visual processing capacity by making visual/diagrammatic representations of all sorts of things that are not naturally visible.
The result is that while you can still (at least arguably) detect our origins in the savannah in our current behavior/thinking, it shows up in marginal deviations from the SSSM and its variants.
I’d be interested to read your thoughts on “overweighting”. From the quotes I don’t know whether you are denying that people place more weight on extreme low-prob events than expected utility would suggest (I disagree) or that doing so should be called “overweighting” (I agree).
Comments on this entry are closed.