I realize the question is broad and could be interpreted in a few different ways so I’ll attempt to narrow it down:
Compared to say, computers, human minds are not particularly good at number crunching. Compared to other apes, we don’t seem to have particularly good hand to eye coordination. Yet there are undeniably tasks or aspect of the mind where we do better than other animals and computers.
For example, the human mind seems to be unusually good at abstracting away random noise to have insight into underlying principles. For example, other animals will readily recognize that something can fall but they’ve never abstracted away the specifics of different situations to learn about gravity itself, seen its relation to other phenomena and then used that general knowledge of gravity to do something unprecedented like a gravity turn in space.
Are there other things at which the human mind is unusually good? Anything else to say about the human mind’s ability to get insights?
The human mind is unusually good at pattern recognition. That translates into everything from facial recognition to humor to creating new music. Computers are getting better at some aspects of that but it is surprisingly difficult to make a machine that is even remotely as good at pattern recognition as the human brain.
I agree with Darren Garrison. Our precision grip (between thumb and fingers) is excellent. And our vision matches it. We may not be able to catch a bird in flight (like Hylobates), but we have very good coordination for most tasks.
To give some examples–the ability to take an arbitrarily chosen series of sounds and use them for encoding and transmitting abstract thoughts. The ability to accomplish the same using an arbitrarily chosen series of visual squiggles. The ability to make multi-step plans, including deep into the indeterminate future.
I’m going to agree with pattern recognition. particularly too, with certain patterns. For example, we can recognized words in music played backward, where no such words actually exist. (I recall a radio demonstration where they played a song backward and it sounded like noise. Then the host said "it sounds like they said ‘the girl with a monkey on her head’ " Once that was suggested, you could sort of make out that garbled phrase in the noise.
Similarly, we see eyes and faces in the weirdest places (and Jesus on our toast) because the mind is particularly designed to recognize the threat of hidden eyes of predators watching for us. Plus, we can tell even identical twins apart if we are familiar with them, since our social structure depends on distinguishing people - by face.
All right then, how about “Misidentifying an objective question as a subjective question, with consequent junior modding about mis-posting into the wrong forum.”?
ETA: more value-added contribution: what are infant humans good at, beyond reflexive physical actions like rooting? Visual pattern recognition (face recognition, apparently), sound recognition (heartbeats and voices), and scent recognition. And even as our cognitive and perceptual abilities grow along with the rest of us, our visual pattern recognition mechanisms continue to improve. Reading, for instance, is visual pattern recognition in the service of non-spoken communications.
It seems to be strong enough that it’s often overactive. To take one notable example: John Nash. The same pattern recognition abilities which enabled him to be a good mathematician also got him into deep delusions. Yet the delusions he had were partly credible; It is plausible that a mathematician would be enlisted by an intelligence agency to decipher codes.
A lesser version of that can be seen in paranoia, various forms of mental illness or political persuasions which can be guessed at but not mentioned in this forum. Someone may have bad experiences then their “threat recognition/prediction algorithms” get overactive.
Which leads me to think that there may be a complementary ability to pattern recognition, some kind of falsificationist filter for lack of a better term.
Pattern recognition is the foundation of being able to generalize. To create categories where categories did not previously exist, and use them to make sense of the world. Pattern recognition allows the human mind to create an hypothesis. A computer, in the form that we know them, can assess data, and so if the data is available it can test the hypothesis. But they aren’t as of yet adept at creating hypotheses. Pattern recognition allows the human mind to formulate a study sufficient to gather the data necessary to test the hypothesis, as well: out of many possible ways to go about it, some of them are more elegant, appealing to us, as we visualize the “shape” of how we would go about it were we to do it this way. Computers aren’t as of yet good at developing a study proposal.
Indeed. We’re qualitatively better at language than any other species on Earth. The fact other species have language doesn’t remove how special humans are at language: We’re the only species with complex, recursive grammars.
Humans are extremely good at finding patterns and forming explanatory hypotheses. However, our false positive / false negative calibration is heavily skewed toward false positives. In other words, we are “tuned” to be extremely sensitive so that we never miss possible patterns and associated cause-effect relationships, but the price is that we often read too much into random noise.
The classic example is walking through the forest at night, when we nervously interpret every rustle in the trees as a potential threat. And the evolutionary explanation is obvious: thousands of false positives cost us very little (it was the wind, not a werewolf), but one false negative may kill us (it was a werewolf, not the wind).
One result of this tendency was that in the pre-scientific world, our culture was full of a bunch of useless crap, with a few bits of useful knowledge hidden in there. “Folk medicine” is a great example. The problem with folk medicine is that any time somebody got well after taking a purported remedy, a cause-effect relationship was assumed. But the relationship was never tested rigorously, and most of the time it did not exist, since people usually get well without any treatment.
That’s why the invention of rigorous scientific methodology has led to such dramatic advances for our species. We were always naturally extremely imaginative at inventing thousands of hypotheses to explain our world, but in the pre-scientific world we were terrible at testing those hypotheses to figure out which ones were right.
One example I saw, of a psychological test where rats outperformed humans: The subject was given an apparatus with two buttons and two lights. The subject could press either button, and then a light would come on. If the button pressed matched the light, then the subject would get a reward (a food treat for the rat, but probably something more abstract for the humans).
The humans kept on finding patterns in the lights, and coming up with rules like “when it’s the left light three times in a row, the next one will be right”, or “when I push the button quickly, the right light comes on, and if it’s slow, it’s the left light”. But really, there was no pattern at all: It was just random, with the left light coming on 80% of the time, and the right light 20%. And while the humans were messing around with trying to find a pattern where there wasn’t one, the rats very quickly learned to just always press the left button, which is the best possible strategy.
What if they tried that with children or infants who don’t have the culturalized expectation that something as artificial as a rigged lighting system is obviously meant to test them of something unusual? And even that rule (3x left, 1x right) is 75%, right, still pretty close? Maybe the “abstract reward” was less incentivizing than our drive to find patterns, especially when we believe the pattern is from someone who’s trying to deliberately fool us.
Devise a similar test with something that SEEMS random (maybe like a pachinko machine?) with similar probabilities and rewards, and I wonder if the rats would still do better…
That sounds suspiciously like the ultimate classic human pattern recognition fail, the “foolproof system” for gambling. People continuously make patterns out of totally random data in the belief it will make them rich - which is better than a food pellet.
It takes trying to learn another language in adulthood, to realize just how Herculean a task natural language is, and yet how effortless it feels once you’ve spoken a language for a while.
IMO it’s much more amazing than being a world-class musician, but basically everyone can do it, so it earns zero credit.
