4 ms·
>> Mammals rarely engage in repeated stereotypical behaviour when presented with a task wherein they cannot directly obtain their goal, but will change their be
by YeGoblynQueenne 6y ago
>> Mammals rarely engage in repeated stereotypical behaviour when presented with
a task wherein they cannot directly obtain their goal, but will change their
behaviour and attempt different strategies. This could provide one possible
definition of intelligence in animals: the more complex the improvised
strategy, the more intelligent the animal. Other behaviours can also be used
as markers of intelligence and there are gradations in intelligence.
This is very tempting to accept. I have no idea how accurate it is (not a
specialist in the subject) but it certainly rings true to my experience.
I might have told the story of the Light Fixture of Death on HN before.
There's a light fixture where I stay in the summers, shaped like a quarter
sphere, on the wall inside the living room. In the summer, large, orange
hornets that have their nest in the chimney start coming out of the fireplace.
In the nighttime, if the Light Fixture of Death is on, they will, invariably,
draw a, well, beeline to it, stand on its edge, slide on its smooth interior and
immolate themselves onto the (incadescent) light bulb with a horrible sizzling
sound. The fixture has to be changed every once in a while because it fills up
with dead hornets. There's a black imprint above it, like you'd expect to see
over a candle or oil lamp, because of all the smoke of burning insects that
emanates from it.
There is no exception to this behaviour. As soon as a hornet enters the living
room via the fireplace, it will always, always fly to the Light Fixture of
Death and burn itself to death. If they come in from another side, say the
balcony doors, then they often miss it.
I've started to think that this is some kind of hallmark of insect
intelligence: they have a set of behaviours that are adaptive in the sense
that they can cope with the variations in an environment but that are also
limited in that they only work in specific environments- and outside of those
environments, an insect's very survival is basically down to luck. Again, I'm
no expert on this so I'm possibly talking bullshit, but I can't imagine the
same kind of stereotypical behaviour in cats or dogs. I mean, I see dead
strays on the edge of the road all the time, but then again I see so many
strays that are not hit by cars. There must be some variation in their
behaviours. Maybe not.
It's tempting also to see insect behaviour as the result of some kind of
million-years' optimisation process, not unlike gradient-based optimisation in
modern statistical machine learning. But that's stretching things way too far.